Unveiling the Social Fabric of Academia: Using SNA to Map Research Collaboration
Collaboration patterns of researchers using Social Network Analysis approach
This paper utilizes Social Network Analysis (SNA) to investigate research collaboration patterns within the Faculty of Electrical Engineering (FKE) at Universiti Teknologi Malaysia (UTM). By analyzing 1446 publications and 168 researchers, the study identifies influential nodes and finds that while internal department cohesion exists, inter-departmental collaboration remains significantly low.
TL;DR
Is individual productivity the only measure of a researcher's value? This paper argues "No." By applying Social Network Analysis (SNA) to the Faculty of Electrical Engineering at UTM, researchers demonstrated that the structural position of a faculty member—their role as a broker or a hub—is just as critical to institutional success as their h-index. The study reveals a high level of departmental "siloing" and identifies the specific individuals who hold the network together.
Problem: The "Lone Wolf" Fallacy in Metrics
Standard Key Performance Indicators (KPIs) in universities often treat researchers as isolated islands, measuring success through publication counts and JIF (Journal Impact Factor). However, science is inherently collaborative. Prior work often ignores the relational aspect of research:
- Invisibility of Brokers: A researcher might have a lower h-index but serve as the only bridge between two isolated departments.
- Fragmented Strategy: Without mapping the network, administrators cannot see "knowledge silos" where departments fail to communicate.
Methodology: Mapping the Co-authorship Topology
The researchers moved beyond simple list-counting by constructing an Adjacency Matrix based on co-authorship data from Scopus (2011-2015). They utilized several sophisticated SNA metrics:
- Degree Centrality: Popularity and communication activity.
- Betweenness Centrality: The "Brokerage" score—who controls the flow of information between groups?
- Eigenvector Centrality: Not just how many people you know, but how influential your connections are.
Figure 1: The visualization of FKE's network highlights the clustering within departments and the scarcity of cross-departmental links.
Key Insights from the Data
The analysis of 168 researchers yielded startling results regarding the density of collaboration:
- The Silo Effect: The overall faculty density was 4%, but individual departments like CMED reached 18.7%. This suggests that while researchers work well with immediate colleagues, they rarely cross departmental boundaries.
- The Power Players: Researchers like A118 and A49 were highlighted not just for their output, but for their high Betweenness Centrality, identifying them as the essential "glue" of the faculty.
Figure 2: Visual comparison showing how CMED develops dense internal relationships compared to more fragmented departments like ECE.
Critical Analysis & Conclusion
Takeaway
The study proves that SNA is a powerful diagnostic tool for academic leadership. It allows for the identification of "Reliable" researchers (via Tie-Strength) and "Influential" researchers (via Centralities) who could lead multi-disciplinary grants.
Limitations
A notable limitation is the data cleaning process: the authors excluded researchers with zero h-index or zero collaborations. While this focuses on "active" contributors, it may create a survivor bias, potentially masking the very isolation problems the study seeks to highlight.
Future Outlook
As research becomes increasingly globalized, moving this SNA approach from a single faculty to a whole-university or inter-university level is the next frontier. Understanding how these networks evolve over time (Temporal SNA) will be key to predicting which research groups will become the SOTA leaders of tomorrow.