The association between Twitter content and suicide
Bibliographic record
Abstract
OBJECTIVE: A growing body of research has established that specific elements of suicide-related news reporting can be associated with increased or decreased subsequent suicide rates. This has not been systematically investigated for social media. The aim of this study was to identify associations between specific social media content and suicide deaths. METHODS: = 787) geolocated to Toronto, Canada and originating from the highest level influencers over a 1-year period (July 2015 to June 2016) were coded for general, putatively harmful and putatively protective content. Multivariable logistic regression was used to examine whether tweet characteristics were associated with increases or decreases in suicide deaths in Toronto in the 7 days after posting, compared with a 7-day control window. RESULTS: Elements independently associated with increased subsequent suicide counts were tweets about the suicide of a local newspaper reporter (OR = 5.27, 95% CI = [1.27, 21.99]), 'other' social causes of suicide (e.g. cultural, relational, legal problems; OR = 2.39, 95% CI = [1.17, 4.86]), advocacy efforts (OR = 2.34, 95% CI = [1.48, 3.70]) and suicide death (OR = 1.52, 95% CI = [1.07, 2.15]). Elements most strongly independently associated with decreased subsequent suicides were tweets about murder suicides (OR = 0.02, 95% CI = [0.002, 0.17]) and suicide in first responders (OR = 0.17, 95% CI = [0.05, 0.52]). CONCLUSIONS: These findings largely comport with the theory of suicide contagion and associations observed with traditional news media. They specifically suggest that tweets describing suicide deaths and/or sensationalized news stories may be harmful while those that present suicide as undesirable, tragic and/or preventable may be helpful. These results suggest that social media is both an important exposure and potential avenue for intervention.
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 | 0.000 | 0.005 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.004 | 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".