Harnessing the Potential of Google Searches for Understanding Dynamics of Intimate Partner Violence Before and After the COVID-19 Outbreak
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Abstract Most social phenomena are inherently complex and hard to measure, often due to under-reporting, stigma, social desirability bias, and rapidly changing external circumstances. This is for instance the case of Intimate Partner Violence (IPV), a highly-prevalent social phenomenon which has drastically risen in the wake of the COVID-19 pandemic. This paper explores whether big data—an increasingly common tool to track, nowcast, and forecast social phenomena in close-to-real time—might help track and understand IPV dynamics. We leverage online data from Google Trends to explore whether online searches might help reach “hard-to-reach” populations such as victims of IPV using Italy as a case-study. We ask the following questions: Can digital traces help predict instances of IPV—both potential threat and actual violent cases—in Italy? Is their predictive power weaker or stronger in the aftermath of crises such as COVID-19? Our results suggest that online searches using selected keywords measuring different facets of IPV are a powerful tool to track potential threats of IPV before and during global-level crises such as the current COVID-19 pandemic, with stronger predictive power post outbreaks. Conversely, online searches help predict actual violence only in post-outbreak scenarios. Our findings, validated by a Facebook survey, also highlight the important role that socioeconomic status (SES) plays in shaping online search behavior, thus shedding new light on the role played by third-level digital divides in determining the predictive power of digital traces. More specifically, they suggest that forecasting might be more reliable among high-SES population strata.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it