Methods · History, human story & context · Short guide

Reading strategy selection

Decide skim, scan, or deep-read for a given source and purpose, so not every PDF gets the same expensive attention.

When to use this: Before you start studying
Learning impact: Clear help, modest gain
Learning evidence: Credible with caveats
Where you are with the topic: Meeting the idea
How often to use it: Worth repeating often

When to use it

Every PDF gets the same highlighter ritual, and you are drowning.
You need one fact from a paper, not mastery of the whole thing.
You finished a "deep read" and cannot say what decision it was for.

What to ask AI

Type something like: "Source: [ ]. My purpose: [ ]. Recommend skim, scan, or deep-read with a time box and a required output. After I finish, check whether my output matches the mode, no content lecture unless I deep-read."

Insist on an output that proves the mode.

Keep the hard part on your side: you produce the output before asking for a lecture.

How it helps

Purpose chooses mode.

Skim maps structure; scan finds a target; deep-read earns understanding you will use.

Matching attention to goal is the metacognitive move, not treating every page as sacred.

How to run it

Name the source and your purpose in one sentence each.
Accept a recommended mode, time box, and required output.
Do that mode only, do not secretly deep-read while claiming a skim.
Show your output and ask whether it matches the mode.
Only then ask for explanation if deep-read was justified.

Example exchange

You

I have a 28-page methods paper on survey weighting. I only need to decide whether their nonresponse adjustment applies to my dataset. Should I skim, scan, or deep-read, and what should I walk away with?

Scan for about 25 minutes. Walk away with the exact nonresponse procedure in three bullets, plus one sentence on the inputs it requires. Skip a literature review.

You

Here are my three bullets and the input list.

That matches a scan. You do not need a deep read of the lit review for this decision.

Copyable prompt

Source: [ ]. My purpose: [ ].
Recommend skim/scan/deep with a time box and required output. After I finish,
check whether my output matches the mode: no content lecture unless I deep-read.

The Tell

Here is how you know this method has flipped: every paper still gets the same highlighter ritual.

If you cannot name the mode and the output you owed yourself, you did not choose a strategy, you performed reading.

Principle evidence

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

The underlying learning idea is rated moderate. Matching reading mode to purpose is well motivated in metacognitive reading research; effects depend on actually producing a checkable output.

AI delivery evidence

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

AI can recommend modes fluently. Whether you follow the time box and output is what determines benefit, that part is untested as an AI product claim.