Method 2 min read
What an evidence-led reading of a geopolitical event looks like
In plain words: how to look at a news event by checking where each claim came from, instead of trusting the loudest headline.
Most coverage starts from a conclusion and collects support. An evidence-led reading starts from the record and sees how far it goes.
When a border incident, a sanctions announcement or a summit communiqué breaks, the first hours produce a great deal of text and very little that can be checked. Headlines arrive with adjectives attached. Analysts and commentators supply a frame before the facts have settled. By the time the details are clear, the frame has usually hardened into the story.
An evidence-led reading reverses the order. It begins with what is on the record: who said what, in which document or broadcast, at what time, and who else has said the same. The frame comes last, and it is held loosely.
Start with the claim, not the story
A story is a bundle of claims. “Talks resumed on Tuesday” is one claim. “Talks resumed after pressure from a neighbouring state” is two: the first can be dated and sourced, the second is an attribution of motive that needs its own support. Breaking a dispatch into separate claims is the first discipline, because each claim can then be traced, scored and challenged on its own terms.
Trace every claim to its origin
For each claim, ask where it first appeared. Many reports that look independent are downstream of one statement, one wire item or one anonymous briefing. Tracing to the origin shows how many separate pieces of reporting stand behind a claim, which is usually fewer than the number of headlines.
- Record the first appearance of the claim, with date and time in one fixed time zone.
- Note whether the origin is a primary document, a named person, an unnamed person or an earlier report.
- Mark every later report as a repeat of the origin or as new reporting.
Show the score and what went into it
Confidence is useful only when the reader can see how it was reached. A bare number invites trust or distrust and teaches nothing. A score that lists its inputs, such as the reliability of the source, the amount of independent corroboration, how recent the reporting is and whether primary documents back it, can be argued with. That is the point. A reader who disagrees can see which input they would weigh differently.
Keep description and conclusion apart
The record can say that three outlets reported a figure, that two of them cite the same ministry statement, and that the statement itself has not been published. It does not need to say whether the figure is right. Readers who must decide, whether to brief a minister, file a report or change a plan, are better served by an exact account of the state of the evidence than by a verdict they cannot inspect.
Describe what the evidence shows. Leave the conclusion to the person who has to act on it.
A short checklist
- 01Split the story into claims.
- 02Find the first appearance of each claim.
- 03Count independent origins, not headlines.
- 04Score with named inputs and show them.
- 05Date everything, and keep corrections visible.
None of this makes an analyst faster at producing an opinion. It makes the opinion easier to defend, easier to correct and easier for someone else to pick up. That is the standard we build 7Analytica to support.