The Moderation of Contentious Content on Twitter
Bibliographic record
Abstract
Retweeting posts is Twitter's most important feature, playing a vital role in enabling the platform to be a virtual town hall that fosters timely discussions. This attribute has been instrumental in drawing a younger, wealthier, and more educated user-base, distinguishing Twitter from its competitors. We were motivated by the observation that the retweet count on popular tweets diminishes over time. In particular, this reduction is greater for contentious tweets. Since, retweets represent endorsements, it is pertinent to understand how self-moderation and platform moderation play a role in their retractions. \n \nWe collected our own datasets and tracked various reasons for retweet loss over time. Leveraging Kaggle datasets, we trained models to predict which tweets would see a significant decrease in retweets; the model's performance extended to previously unseen datasets. Additionally, we proposed an algorithm to estimate the timeline of retweet loss and explored factors that contribute to individual unretweeting behaviour. Finally, our data collection period coincided with the volatile phase on Twitter following Elon Musk's acquisition. As a result, we were able to observe the impact of various changes in platform moderation through our analysis.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.013 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".