How we work

Editorial policy

How the What's happening tracker decides what to call confirmed, how it treats the people and organizations it names, and where our safety guidance and product picks come from.

Every card makes two separate judgments

Most events clearly happened. What people argue about is the cause. So a card never gives one verdict for both.

Judgment 1

What happened

One status for each specific claim. When only part of a claim holds, we split it: the outages in three counties can be Confirmed while the statewide claim stays Unconfirmed.

Judgment 2

The cause

Always a sentence that names who is speaking, with a label for how far that explanation has been checked.

Then the action

For your household

The official instruction that applies, with its place and expiry, or practical steps that hold whatever the cause turns out to be.

The July 2024 CrowdStrike outage shows why the two are kept apart. The outage was confirmed within hours. The widely shared suspicion of a cyberattack was not: the cause was a defective software update.

What happened: the three statuses

Each claim on a card carries exactly one of these. The word is always shown, and the color only supports it.

Confirmed
Records or several independent reports show it.Two outlets repeating one wire story, or three articles repeating one company statement, count as one source.
Unconfirmed
There is not enough evidence yet.It does not mean false. When the event itself is unconfirmed, the card has no cause section at all.
Contradicted
Records directly disprove the precise claim.We quote the exact claim, date it and link the records. A claim that is only unlikely or unsupported stays Unconfirmed.
  • For a claim about the future, such as "a 90-day outage is coming", we check what can be verified today, for example whether any agency issued that warning.
  • The affected area, the current official instruction, its expiry and time-stamped counts sit at the top of the card.
  • Each event is marked active or resolved, and each card shows the time we last checked it.

The cause: the four labels

The cause is always written as an attributed sentence, such as "The FAA says excess network traffic disrupted communications." The label says how far that explanation has been checked.

Official account
The affected organization or agency's own explanation.Nobody else has checked it yet. A preliminary explanation is described as preliminary.
Independently confirmed
A source that examined the evidence for that specific conclusion confirms it.An agency title or a newspaper byline alone is not enough.
Disputed
A credible source has documented a contradiction of a specific claim.An online theory never makes a cause Disputed.
Cause unknown
No source has given an explanation that we can attribute.The card says what would settle the question, such as a final investigation report.

How we weigh sources

Every source on a card is shown with its role. Roles carry weight in this order, from most to least.

  1. Independent investigators and regulatorsFor example the NTSB, CISA, the FBI and national cyber agencies, when they have examined the evidence.
  2. Independent researchers and original reportingPeople who looked at the evidence themselves, and newsrooms that did their own reporting.
  3. The affected organization, about itselfA company or agency explaining its own failure. This is reported as its account.
  4. Governments that are a party to a conflictA government's statement about who attacked it is presented as that government's claim.
  5. Social postsNever a source on their own. A post can prompt us to look, and it cannot confirm anything.

Claims about named people and organizations

A card can say that a company's explanation has not been independently confirmed. That is a statement about the evidence, and it is never a suggestion that anyone is being dishonest.

What we always do

  • Attribute every explanation to its source, in the same plain words for a company and an agency: "Microsoft says", "the FAA says".
  • Quote the exact claim, date it and link the evidence.
  • Ask the person or organization for comment before we publish a claim that they did something wrong.
  • Say what is and is not independently confirmed, and what would settle the question.
  • Keep evidence that an attack happened separate from a claim about who carried it out.

What we never do

  • Say or imply that a person or an organization is lying.
  • Use a warning-style badge that suggests deception.
  • Mark a claim Contradicted because it seems unlikely. Records have to disprove it directly.
  • Treat an online theory as a reason to call a cause Disputed.
  • Name a cause, including AI, that no named source confirms.

What publishes automatically and what waits for a person

Speed matters when something is happening, so verified facts publish without waiting. Judgments about responsibility wait for a person. The How we use AI page describes the automated steps.

Automatic

Publishes after verification

  • A verified account of what an agency issued or what records show.
  • The official instruction that applies, with its place and expiry.

When no relevant official instruction exists, the card leaves generic advice out. The verified event publishes while any attribution is still waiting for review.

Human review

Waits for a person to approve

  • Any statement that a named person, company, contractor or state is responsible, even when it is called preliminary.
  • A Disputed cause.
  • A Contradicted status on a claim about a named person or organization.
  • Official instructions that conflict with each other.
  • New medical or evacuation advice.

Blocked

Never publishes automatically

  • Safety instructions that do not come from an official source.
  • Casualty estimates.
  • Nationwide threats sourced from social posts.

Where safety guidance comes from

Named authorities, with the date we read them

Every safety instruction and every number, such as gallons of water per person, comes from a named authority: Ready.gov and FEMA, the CDC, the American Red Cross, the National Weather Service, the EPA, the FDA, the USDA, CISA, a utility or a state emergency agency.

Each guide lists its sources with the date we read them, and shows the date the guide was reviewed.

Medical questions

Medical content comes only from the CDC, the Red Cross, the FDA, Ready.gov or the NIH, and is presented as theirs. For medication, refrigerated medicines and powered medical equipment, we point you to your prescriber and your equipment supplier.

How product recommendations are chosen

From published specifications

We choose products from the manufacturer's published specifications and from official guidance on what a household needs. PrepHabit does not test products hands-on and never describes a pick as tested.

Prices come from Amazon

A price is shown with the time it was checked. When we do not have a recent price, the page shows no price and sends you to Amazon to see the current one.

No promises we cannot back

We do not call any kit complete, promise savings, or say a product is safe for a purpose its manufacturer does not state. When the evidence says you need nothing new, the page says so.

Found something wrong? The corrections page explains how we fix mistakes and how to report one.