[Paper] Does social identity matter in software engineering? Assessing the case of research software engineers

Source: arXiv - 2604.25831v1

Overview

The paper investigates whether social identity—the sense of belonging to a professional group—affects the day‑to‑day lives of Research Software Engineers (RSEs). By analysing thousands of online posts, blog articles, and survey responses, the authors show that a shared RSE identity is emerging and that it has measurable consequences for job satisfaction, collaboration, and career development.

Key Contributions

• Empirical evidence of a collective RSE identity across multiple online platforms (Twitter, blogs, forums).
• Mixed‑methods pipeline that blends computational linguistics (topic modeling, sentiment analysis) with classic inferential statistics.
• Quantified link between identity strength and professional wellbeing (e.g., higher self‑reported job satisfaction and lower burnout among strongly identified RSEs).
• Interdisciplinary framework that adapts social‑psychology constructs (social identity theory, in‑group/out‑group dynamics) for software‑engineering research.
• Open dataset and analysis scripts (≈28 k posts, 1.7 k blogs, 381 survey responses) released for reproducibility and further study.

Methodology

1. Data Collection – The team harvested public social‑media posts (Twitter, Mastodon, Reddit) containing RSE‑related hashtags, scraped blog posts from known RSE community sites, and ran an online survey targeting self‑identified RSEs.
2. Computational Linguistic Analysis – Using Python’s gensim and spaCy, they performed:
   • Topic modeling (LDA) to surface recurring themes (e.g., “career pathways”, “tool sharing”, “community events”).
   • Sentiment & emotion detection (VADER, NRC lexicon) to gauge affective tone.
   • Identity‑keyword extraction (e.g., “we”, “our community”, “RSE”) to compute an “identity salience score”.
3. Statistical Modeling – Survey items measured identity strength (adapted from the Collective Self‑Esteem Scale) and wellbeing (Job Satisfaction Scale, Maslach Burnout Inventory). Linear regression and mediation analysis tested whether identity salience predicts wellbeing outcomes, controlling for experience, role, and institution type.
4. Triangulation – Qualitative coding of a random sample of posts and blog excerpts validated the automated findings and surfaced nuanced narratives (e.g., mentorship stories, feelings of professional legitimacy).

Results & Findings

Overall, the study demonstrates that social identity is not a peripheral academic curiosity; it is a measurable factor shaping RSEs’ professional experience.

Practical Implications

• Team Leads & Project Managers: Foster an explicit RSE community within research labs (e.g., regular “RSE round‑tables”, shared Slack channels). A clear group identity can improve morale and reduce turnover.
• Universities & Research Institutes: Recognize RSEs as a distinct career track in hiring, promotion, and funding policies. Formal titles and dedicated career ladders reinforce the identity that the study shows to be beneficial.
• Tool Builders & Platform Vendors: Tailor documentation, tutorials, and support forums to the RSE audience. Highlight community‑driven best practices to strengthen the sense of belonging.
• Open‑Source Projects: Encourage contributions from RSEs by labeling issues as “RSE‑friendly” and providing mentorship pathways—this aligns with the identity‑driven motivation uncovered in the paper.
• Professional Societies: Use the provided dataset and analysis scripts to benchmark community health over time, similar to how developer ecosystems track “developer happiness”.

In short, building and nurturing a collective RSE identity can be a low‑cost lever for higher productivity, better collaboration, and reduced burnout.

Limitations & Future Work

• Self‑selection bias: Participants who responded to the survey or posted publicly may already be more community‑oriented than the broader RSE population.
• Cross‑sectional design: The study captures a snapshot; longitudinal data would be needed to prove causality between identity formation and wellbeing.
• Language & regional scope: Data were primarily English‑language and Western‑centric, limiting generalizability to non‑English‑speaking RSE communities.
• Future directions suggested by the authors include:
  1. Longitudinal tracking of identity evolution as RSE roles mature.
  2. Experimental interventions (e.g., identity‑building workshops) to test causal impact on performance.
  3. Expanding the analysis to other software‑engineering sub‑communities (e.g., DevOps, data‑engineers) for comparative insights.

Authors

• Chukwudi Uwasomba
• Tamara Lopez
• Melanie Langer
• Helen Sharp
• Michel Wermelinger
• Caroline Jay
• Mark Levine
• Bashar Nuseibeh

Paper Information

• arXiv ID: 2604.25831v1
• Categories: cs.SE
• Published: April 28, 2026
• PDF: Download PDF