Methods · Working with the tool itself · Short guide

Source triangulation

Confirm a claim across three independent sources, so one confident paragraph stops being enough.

When to use this: When checking yourself
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Checking readiness
How often to use it: Worth repeating often

When to use it

One confident paragraph is doing too much work in your notes.
A claim will enter an essay, lab report, or presentation.
Two AI chats agreed, and you are tempted to treat that as proof.

What to ask AI

Type something like: “Claim: [ ]. Help me plan triangulation: what kinds of independent sources would count. Do not invent citations.”

After you open three real sources, paste what you found and ask for a map of agreement and residual doubt.

Keep the hard part on your side: you open the sources. Three model paragraphs are not three sources.

How it helps

Fluency from one place feels like confirmation.

It is not.

Triangulation asks for independent corroboration: different authors, datasets, or methods that are not echoing each other.

Agreement across models can still be shared training residue. Agreement across opened sources is a different kind of signal.

How to run it

State the claim in one sentence you could falsify.
Ask what kinds of independent sources would count for this claim type.
Open three non-echoing sources yourself. Prefer primary or high-quality secondary over chat summaries.
Map: what all three support, what only one asserts, what remains disputed.
Update your confidence. Write the residual doubt in your notes.

Example exchange

You

I want to check this claim: “Spaced practice reliably beats massed practice for long-term retention in classroom studies.” Help me plan how to triangulate it, and please don’t invent citations.

Look for three independent kinds of source: a review or handbook chapter, one primary classroom experiment, and one critical or boundary-condition paper. Independence means different authors and datasets, not three chatbots.

You

Okay. I opened three real sources and pasted the titles and key sentences here.

All three support a spacing advantage under common conditions. What’s still open is effect size and schedule details. Cite what you opened, not me.

Copyable prompt

Claim: [ ].
Help me plan triangulation: what kinds of independent sources would count. Do
not invent citations. After I paste three real sources I opened, help me map
agreement and residual doubt.

The Tell

Here is how you know this method has flipped: your footnotes outnumber the PDFs you actually opened.

Triangulation that never leaves the chat is theatre.

If you cannot point to three opened sources, you have not triangulated.

Principle evidence

Strength of the underlying learning idea, not a claim about AI products.

The underlying learning idea is rated moderate. Multi-source corroboration is a core information-literacy move: independent confirmations beat single-source fluency. The method asks you to open three non-echoing sources, not to trust three model paragraphs.

AI delivery evidence

Whether an AI tutor delivers this method well is a separate question.

AI can suggest where to look. That help is delivery, and unproven as a learning intervention here. The principle lives in your corroboration work outside the chat.