PlatformGovernance as Decision Support: A Governance RiskIndex for Twitter/X During the 2024 U.S. Presidential Campaign

Abdullah Önden

Authors

Abdullah Önden Author

Abstract

Online platforms have emerged as essential infrastructure for democratic deliberation, crisis communication, and public healthmessaging.Nonetheless, governance of these spaces remains largely retrospective and principle-based, without concrete evaluative criteria.This article operationalizes six key principle-based elements of platform governance into concrete clause components with quantifiable triggers,to support proactive application. We develop a composite Governance RiskIndex (GRI), integrating empirically defined decision thresholds for four risk components: dispersion,drift, inequality, and toxicity. We estimate toxicity levels from a stratified subsample of the dataset labeled using the Perspective API, aggregated to the daily level. Our results show that toxicity is positively correlated with engagement (r= 0.52,p< .001), as replicated in our data (r= 0.49,p< .001),in line with algorithmic amplification dynamics thatplatform governance must account for. We further find that decision thresholds differ meaningfully between clause components and risk components. A contextual benchmark against 2015–2016 multi-platform baselines reveals that the 2024 Twitter/X environment exhibits substantially different statistical properties in engagement dispersion and toxicity prevalence, highlighting the need for context-dependent calibration of platform governance thresholds. Over a 30-day out-of-sample holdout period, validated exclusively against independently verified external events, the GRI classified governance-event days under retrospective validation with 85.1% accuracy. Under a supplementary labeling scheme that incorporates platform-internal anomaly criteria, accuracy reaches 90.0%, representing an improvement of about 13 percentage points over single-metric baselines, highlighting the advantages of multivariate, multi-faceted operationalization in proactive platform governance.

Author Biography

Abdullah Önden

Department of Computer Engineering, Istanbul University, Faculty of Computer and Information Technologies,Istanbul, Türkiye

References

https://doi.org/10.3390/systems14020216.

[23]Suoniemi, S., Meyer-Waarden, L., Munzel, A., Zablah, A. R., & Straub, D. (2020). Big data and firm performance: The

roles of market-directed capabilities and businessstrategy.Information & Management, 57(7), 103365.

https://doi.org/10.1016/j.im.2020.103365.

[24]Shiells, K., Di Cara, N., Skatova, A., Davis, O. S., Haworth, C. M., Skinner, A. L., ... & Boyd, A. (2022). Participant

acceptability of digital footprint data collectionstrategies: an exemplar approach to participant engagement and involvement in the ALSPAC birth cohort study.International Journal of Population Data Science, 5(3), 1728.https://doi.org/10.23889/ijpds.v5i3.1728.

[25]Jayasuriya, D. D., Ayaz, M., & Williams, M. (2023).The use of digital footprints in the US mortgage market.

Accounting & Finance, 63(1), 353-401.https://doi.org/10.1111/acfi.12946.

[26]Faruq, M. O. (2024). Vendor risk management in cloud-centric architectures: A systematic review of soc2, Fedramp,

and ISO 27001 practices.International Journal of Business and Economics Insights, 4(1), 01-32.

https://doi.org/10.63125/j64vb122.

[27]Saura, J. R., Palacios-Marqués, D., & Ribeiro-Soriano, D. (2025). Privacy concerns in social media UGCcommunities:

Understanding user behavior sentiments in complex networks: JR Saura et al.Information Systems and e-Business Management, 23(1), 125-145.https://doi.org/10.1007/s10257-023-00631-5.

[28]Iwan-Sojka, D. (2025). The inclusive data governance models for algorithms–a dream of the already convinced or a

realistic way of action?.Information & Communications Technology Law, 34(1), 3-16.

https://doi.org/10.1080/13600834.2024.2406668.

[29]Aldhi, I. F., Suhariadi, F., Rahmawati, E., Supriharyanti, E., Hardaningtyas, D., Sugiarti, R., & Abbas, A. (2025).

Bridging digital gaps in smart city governance: themediating role of managerial digital readiness andthe moderating role of digital leadership.Smart Cities, 8(4), 117.https://doi.org/10.3390/smartcities8040117.

[30]Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions.Advances in Neural

Information Processing Systems, 30.

[31]Önden, A., Kara, K., Önden, İ., Yalçın, G. C., Simic, V., & Pamucar, D. (2024). Exploring the adoptionof the metaverse

and chat generative pre-trained transformer: A single-valued neutrosophic Dombi Bonferroni-based method for the selection of software development strategies.Engineering Applications of Artificial Intelligence, 133, 108378.https://doi.org/10.1016/j.engappai.2024.108378.

[32]Cunneen, M., AnandFinn, R., Friel, R., Tennent, P.,& Brandt, S. (2025). From bones to bytes: anticipating and

addressing the governance challenges of human digital remains and posthumous digital human twins.AI & Society.https://doi.org/10.1007/s00146-025-02514-4.

[33]Jonnala, N. S., Ram Teja, A. V. S., Rajeswari, S. R., Jakeer, S., Dheeraj, A., Bansal, S., ... & Al-Mugren, K. S. (2025).

Leveraging hybrid model for accurate sentiment analysis of Twitter data.Scientific Reports, 15(1), 24438.https://doi.org/10.1038/s41598-025-09794-2.

[34]Torgo, L., & Moniz, N. (2018). News popularity in multiple social media platforms [Dataset].UCI Machine Learning

Repository.https://doi.org/10.24432/C5H029.

[35]Hutto, C., & Gilbert, E. (2014). Vader: A parsimonious rule-based model for sentiment analysis of social media text.

InProceedings of the International AAAI Conference onWeb and Social Media, pp. 216-225.

https://doi.org/10.1609/icwsm.v8i1.14550.

[36]Devlin, J., Chang, M. W., Lee, K., & Toutanova, K.(2019). Bert: Pre-training of deep bidirectional transformers for

language understanding. InProceedings of the 2019 conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 4171-4186.https://doi.org/10.18653/v1/N19-1423.