Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-0121
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-07-13T06:28:32.697577Z
status: published
lens: g_space
sector: health
headline: Leveraging social media footprints for predicting college student anxiety: a machine learning approach
dek: Background Anxiety is one of the most prevalent mental health concerns among college students worldwide, yet traditional assessment methods relying on self-report questionnaires are time-consuming, susceptible to response bias, and difficult to scale. Social media platforms, which students use extensively, generate rich behavioral and linguistic data that may reflect underlying psychological states. This study investigates whether passively collected social media footprints are associated with anxiety scores amo…
gain_title: A Random Forest model trained on linguistic, emotional, cognitive, behavioral and temporal features from Weibo posts predicted Self-Rating Anxiety Scale scores among consenting Chinese college students with R2 0.77 on the test set.
problem_title: (none)
trace_subject: (none)
gain_reading: A Random Forest model trained on linguistic, emotional, cognitive, behavioral and temporal features from Weibo posts predicted Self-Rating Anxiety Scale scores among consenting Chinese college students with R2 0.77 on the test set.
gain_evidence: Public social-media signals were associated with SAS anxiety scores in this consenting, active Weibo-user sample and supported cross-sectional score prediction within the study context
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers surveyed college students in China with the Self-Rating Anxiety Scale and, with informed consent, analyzed their public Weibo posts. Using multi-dimensional features, a Random Forest model predicted anxiety scores within the study sample, achieving the best test performance among four models tested.

This matters as a proof-of-concept for digital phenotyping that could help identify anxiety-related behavioral patterns at scale without relying solely on questionnaires. It remains unclear whether the approach would work outside consenting active Weibo users, translate to clinical outcomes, or maintain performance over time.
limitation: Findings are cross-sectional, limited to consenting active Weibo users in the sample, and show association and score prediction within study context, not clinical diagnosis or longitudinal forecasting.
tag: Evidence-backed gain
key_points: Cross-sectional study surveyed 3,211 students across 11 universities in four provinces of China using the Self-Rating Anxiety Scale, yielding 2,368 valid responses. | With informed consent, 83,968 Weibo posts were collected and 56,060 posts from 1,316 users retained after preprocessing. | Four models compared were Random Forest, XGBoost, LightGBM, and SVR, with SHAP values used for interpretation. | Top contributing features included grade level, professionalism, emotional expression, risk-related language, curiosity index, emotional tone, and word count.
rundown: By July 11 2026, researchers reported a cross-sectional study of 3,211 students across 11 universities in China, collecting 83,968 Weibo posts with consent and retaining 56,060 posts from 1,316 users after preprocessing. They extracted multi-dimensional linguistic, emotional, cognitive, behavioral and temporal features and trained four models to predict Self-Rating Anxiety Scale scores.

The result matters because it shows passively generated public social media data can support transparent, reproducible estimation of anxiety-related patterns in this student population, offering a scalable complement to self-report questionnaires that are time-consuming and prone to bias. Uncertainty remains about generalizability beyond active Weibo users, clinical validity, and whether such prediction would hold longitudinally or in other cultural contexts.
sources:
- peer_reviewed | BMC Psychology | https://doi.org/10.1186/s40359-026-05161-6 | 2026-07-11
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

c460be4b1d3db9b0ab256aab4a20821d5b761a028834c3601442c9e6fe2a7042

previous

0000000000000000000000000000000000000000000000000000000000000000