Provenance
Outputs are tied to specific archived alert events, pre-specified windows, source datasets and documented comparison logic.
SCC Nexus · public safety evidence
UK Emergency Alert Impact Observatory
A public research programme examining measurable effects around UK mobile Emergency Alerts without converting statistical association into causal claims.
Purpose
Measure what happened around an alert, without overstating why.
The Observatory publishes validated, non-sensitive outputs and deliberately separates discovery, sensitivity analysis and causal interpretation. Personal or identifiable safeguarding and health records are outside the public layer.
The programme follows the SCC Nexus discipline of Search · Corroborate · Communicate, with corrections and collaboration routed through hello@sccnexus.co.uk.
Current evidence
The primary historical result remains negative.
- 23 Apr 2023: 3 police-reported personal-injury collisions in the 0–<15 minute window versus a matched-control mean of 4.8125; RR 0.623; empirical p 0.784.
- 7 Sep 2025: 7 versus 5.5625; RR 1.258; empirical p 0.197. Not statistically persuasive on the pre-specified primary endpoint.
- 7 Sep 2025, 0–<5 minutes: RR 1.739; empirical p 0.057. Exploratory only and not replicated in 2023.
- 14 Aug 2026 wildfire case: suitable collision outcome data are still awaited.
2026 operational alert
Two archived broadcasts are retained as distinct timestamps.
GOV.UK records a Welsh-language broadcast to Wales at 18:57 BST and a second alert to England and Wales at 19:01 BST on 14 August 2026. The Observatory keeps these timestamps separate because collapsing them could misclassify short-window exposure.
| Broadcast | Area | Current project status | Source |
|---|---|---|---|
| 14 Aug 2026 · 18:57 BST | Wales | Awaiting suitable outcome data | GOV.UK archive ↗ |
| 14 Aug 2026 · 19:01 BST | England and Wales | Awaiting suitable outcome data | GOV.UK archive ↗ |
Evidence governance
Provenance · integrity · verification · validity
Integrity
The public Observatory is separated from the research layer; sensitive or identifiable health and safeguarding material is excluded from publication.
Verification
Outputs distinguish analysed, non-persuasive, exploratory and unavailable evidence instead of collapsing them into one headline.
Validity
Weather, emergency response, publicity and behavioural changes can coincide with alert timing. Association alone does not establish causation.
Public evidence record updated 31 August 2026
Limitations & contact
What this project will not claim.
A null STATS19 result does not prove no driving disturbance occurred: damage-only incidents and near misses can be absent. An exploratory short-window signal does not override the pre-specified primary endpoint. External news is context, not a finding.
General enquiries, corrections and research collaboration: hello@sccnexus.co.uk.