Methods · History, human story & context

Reading a paper properly

Read a research paper in the order a specialist reads it, and know what to do when you get stuck, which is a skill nobody teaches and everyone is assumed to have.

When to use this: While you practise
Learning impact: Strong change
Learning evidence: Credible with caveats
Where you are with the topic: Building understanding
How often to use it: Worth repeating often

When to use it

A paper feels impossible because you are reading it like a textbook chapter.
You need to triage: is this paper worth a deep read for your question?
You stall at the first unfamiliar term and never reach the claim.

What to ask AI

Type something like: "Here's the abstract and figure captions. What is this paper claiming, what kind of evidence does it offer, and what would I need to already know to read it?"

Ask for a specialist reading order, not front-to-back.

Keep the hard part on your side: you attempt the claim and evidence type before a guided tour of every section.

How it helps

Specialists do not read papers front to back like chapters.

Strategy, claim, evidence type, required background, then selective deep reading, turns an opaque PDF into a decision.

Triage exists so you can abandon papers that do not serve your question.

How to run it

Paste abstract and figure captions first.
Extract claimed contribution and evidence type.
List background you are missing.
Skim results/figures before drowning in methods.
Abandon or deep-read based on your purpose.
Read past one unfamiliar term before stopping.

Example exchange

You

Here's the abstract and the figure captions from a paper. What is it claiming, what kind of evidence is that, and what do I already need to know?

It claims a new regularisation improves calibration on a named benchmark. The evidence is empirical tables and figures, not a proof. You will need the calibration metrics and the baseline models named in figure 1. What is your purpose, replicate, cite, or skim for method?

You

Cite whether the calibration gains look real for my setting.

Then read the figures and the comparison protocol next, and postpone the related-work prose. Paste one figure takeaway when you are done.

Copyable prompt

Here's the abstract and figure captions. What is this paper claiming, what kind
of evidence does it offer, and what would I need to already know to read it?

The Tell

Here is how you know this method has flipped: you still read every paper front to back, every word.

Or you never abandon, triage failed, or you deep-read everything including papers that did not warrant it.

Principle evidence

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

The underlying learning idea is rated moderate. Disciplinary literacy and strategy instruction for scholarly reading have a credible education research base; effects depend on matching strategy to purpose.

AI delivery evidence

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

AI walkthroughs can still lecture section-by-section. Keep it on triage and order; verify any paraphrase of results against the PDF.