A student opens Google.
Types six words.
Clicks the first result.
Reads three paragraphs.
Finds one sentence that looks useful.
Copies it.
Research complete.
No.
Another student opens twenty-seven tabs.
Wikipedia.
Two newspapers.
Four blogs.
A government page.
Three PDFs.
Five AI answers.
A YouTube transcript.
Several pages they no longer remember opening.
Three hours later, they have a great deal of information.
They are not quite sure what they know.
Also no.
These are opposite-looking failures.
One stops far too early.
The other does not know how to stop at all.
Both have the same underlying problem:
search effort is not being governed.
That matters enormously now.
For most of human history, information was expensive to obtain.
Today, information is cheap to retrieve and expensive to govern.
A student can receive thousands of search results in a fraction of a second.
An AI system can produce an answer before the student has finished deciding whether they asked the right question.
A research database can return three thousand papers.
A video platform can suggest another explanation forever.
The bottleneck has moved.
It is no longer simply:
Can I find information?
It is:
Can I search an enormous information environment without being captured by the first plausible answer, wandering indefinitely, mistaking relevance for reliability, or collecting more material than I can turn into knowledge?
That is why Information Foraging belongs in the Top 10 … Skills Worth Learning series.
A useful Wintour House definition is:
Information foraging is the disciplined search for useful information under limited time and attention: define the information need, create informative search moves, follow promising cues, explore enough of the landscape to avoid premature commitment, deepen where the expected value is high, leave low-yield patches, preserve where information came from, and stop when additional searching is unlikely to change the understanding or decision enough to justify the cost.
The word foraging is useful.
Not because students should behave like animals looking for food.
Because search environments contain something remarkably similar to patches.
One website.
One database.
One book.
One expert.
One search-result cluster.
One AI conversation.
One citation network.
One archive.
Each patch has a possible yield.
Entering it costs time.
Reading deeper costs more.
Leaving costs something too because a new patch must be located and understood.
The learner constantly faces a hidden decision:
Stay here?
Or:
Move somewhere else?
Information-foraging theory formalises exactly this sort of navigation problem. In Peter Pirolli’s 2005 Cognitive Science paper, web navigation is modelled partly through cues linked to a user’s information goals and through continuing judgments about the expected benefit of remaining in one information patch versus leaving for another.
This is not merely an internet skill.
A child looking through a textbook is foraging.
A student deciding which paragraph deserves close reading is foraging.
A researcher following references is foraging.
A learner deciding whether to ask the teacher, inspect the glossary, watch an explanation or attempt another question is foraging.
An AI user deciding whether to continue prompting the same conversation or seek an independent source is foraging.
The Wintour House question is therefore deliberately durable:
If a learner became excellent at ten information-foraging operations, which ten would still matter even if search engines, libraries, databases and AI systems changed completely?
Before the Top 10: Finding Something Is Not the Same as Searching Well
Imagine a game.
There are twenty possible objects.
You need to identify one.
Question A:
“Is it the red bicycle?”
If no, you have eliminated one possibility.
Question B:
“Is it something that can move?”
Depending on the set, that question might eliminate half the possibilities.
Both are legitimate questions.
One has potentially much greater information gain.
Developmental researchers studying children’s information search make a useful distinction between whether a search eventually succeeds and whether the actions chosen were efficient and informative. Török, Stanciu and Ruggeri’s 2024 Child Development Perspectives paper argues that efficiency, effectiveness and success should be distinguished: a child can reach the correct answer through luck, while a strategically informative search can occasionally fail because of the particular sequence of outcomes.
That distinction belongs everywhere.
Student A searches:
“plants”.
Student B searches:
“why does increased leaf surface area increase water loss through transpiration”.
The second search has narrowed the uncertainty far more aggressively.
But another day, the first student may happen to click exactly the page they needed.
Lucky success does not make the first strategy better.
Information Foraging therefore evaluates the route, not merely the final click.
1. Learn to Define the Information Gap Before Searching
The weakest search begins:
“I need information about photosynthesis.”
That is not yet an information gap.
What do you already know?
What exactly is missing?
Perhaps:
“I understand that plants manufacture glucose, but I cannot explain why light intensity eventually stops increasing the rate.”
Now there is a target.
Or:
“I need research for my GP essay.”
Too large.
Try:
“I need evidence about whether later school start times actually increase adolescent sleep duration rather than merely delaying morning reporting time.”
Much better.
Or:
“I cannot solve this Mathematics problem.”
Still broad.
Try:
“I know the algebra after the equation is formed, but I cannot tell which quantities in the word problem should become the variables.”
Now the information search can become surgical.
This is the boundary with Top 10 Problem-Framing Skills Worth Learning.
Problem Framing owns the larger question:
What problem are we actually facing?
Information Foraging begins once part of that problem becomes:
What information do I need to reduce this uncertainty?
A useful sentence is:
I already know X. I need Y in order to decide or understand Z.
That sentence prevents enormous amounts of wandering.
It also changes how students use AI.
Weak:
“Teach me electricity.”
Stronger:
“I understand current and voltage separately, but I cannot explain why adding resistance reduces current when voltage is held constant. Help me identify the missing relationship first.”
The second prompt does not merely ask for more information.
It defines the missing information.
Worth learning because: search becomes efficient only when the learner knows which uncertainty the next piece of information is supposed to reduce.
2. Learn to Turn the Gap Into High-Information Search Terms
A good question is not automatically a good search query.
Humans speak in sentences.
Search environments respond to cues.
The learner needs to translate.
Suppose the question is:
“Why do students remember material better when retrieval practice is spaced rather than massed into one session?”
Possible search concepts:
retrieval practice,
spacing effect,
delayed retention,
distributed practice,
testing effect.
Now the student possesses several routes into the literature.
This is not keyword stuffing.
It is conceptual vocabulary.
Searches improve when learners identify:
the central concept,
a synonym,
the relationship,
the population or context,
the type of evidence needed.
A Primary learner may simply learn that searching:
“animal”
is weaker than:
“how do fish breathe underwater gills”.
A Secondary learner can learn to reformulate.
No good results?
Try the scientific term.
Try a broader term.
Try a narrower term.
Remove an unnecessary condition.
Add a date.
Add a jurisdiction.
Add “systematic review”.
Add “official”.
Add the exact phrase from a syllabus.
JC learners should become particularly comfortable with the fact that failed search results can diagnose the query.
Search returned irrelevant papers?
Perhaps your terminology belongs to another field.
Search returned only consumer articles?
Perhaps you have not used the technical construct.
Search returned thousands of papers?
Perhaps the question is still too broad.
This connects to Top 10 Questioning Skills Worth Learning, but the jobs stay separate.
Questioning constructs the epistemic question.
Information Foraging converts that question into moves an information environment can answer.
Worth learning because: the vocabulary used to search determines which part of the information landscape becomes visible.
3. Learn to Read Information Scent Before You Commit Attention
A search-result page contains clues.
Title.
Snippet.
Domain.
Author.
Date.
Journal.
Document type.
Headings.
References.
Language.
A learner can use these cues to estimate:
Is this likely to contain the information I need?
That is information scent.
Pirolli’s information-foraging work uses the term for proximal cues that help connect available navigation choices to a user’s information goals.
But there is a critical Wintour House boundary.
Information scent is not credibility.
A page can smell highly relevant and be completely wrong.
A headline might match your question perfectly.
That tells you:
inspect me.
It does not tell you:
believe me.
Likewise, a dense academic paper can have weak scent for a Primary learner because the learner cannot yet decode its vocabulary, even if the paper is authoritative.
Search skill therefore asks two separate questions.
First:
Is this likely to contain what I need?
Later:
Should I trust what it says?
The second belongs to Verification and Critical Thinking.
This distinction becomes extremely important with AI-generated search summaries.
The AI can produce wonderfully strong scent.
It recognises your wording.
Compresses the apparent answer.
Highlights exactly the concepts you requested.
That makes it attractive.
It does not eliminate the need for source checking.
The more perfectly a result matches the learner’s question, the easier it is to forget that relevance and truth are different axes.
Worth learning because: information scent helps learners allocate attention efficiently, but strong relevance cues should trigger inspection rather than automatic belief.
4. Learn to Explore the Landscape Before Committing to the First Patch
The first result is dangerous.
Not because it is necessarily bad.
Because it becomes an anchor.
A student searches:
“Does homework improve learning?”
First page:
“Homework is essential for success.”
Now every later source is interpreted around that frame.
Another student’s first result says:
“Homework has no benefit.”
Different starting point.
Different landscape.
Information Foraging therefore needs an exploration pass.
Before reading one source deeply, look across several results.
What kinds of sources exist?
Official guidance?
Primary studies?
Systematic reviews?
News coverage?
Educational explainers?
Opposing interpretations?
Different dates?
Different populations?
The goal is not to read everything.
It is to create a rough map.
A professional researcher does this instinctively.
Before falling into one paper, they want to know:
What field am I in?
Which authors recur?
Which terminology recurs?
Is there a review?
Is this a mature literature or a new debate?
Which source appears primary?
Which appears derivative?
The same principle can be taught much earlier.
Primary:
look at three possible pages before choosing one.
Secondary:
scan several results and identify source type.
JC:
conduct a landscape pass before close reading.
This prevents one patch from silently defining the entire problem.
It also reveals something surprisingly valuable:
the information environment itself contains information.
If every result points back to one original study, that matters.
If reputable sources disagree, that matters.
If recent work uses a different term from older work, that matters.
The search landscape is not just a menu.
It is evidence about the state of the question.
Worth learning because: brief exploration reduces first-result lock-in and gives the learner enough of the landscape to decide where deeper attention is worth spending.
5. Learn to Balance Exploration and Exploitation
Search contains a permanent trade-off.
Explore:
look for new patches.
Broaden the landscape.
Find alternatives.
Discover terminology.
Exploit:
stay in one promising patch.
Read deeply.
Follow its references.
Extract details.
Understand the mechanism.
Too much exploration creates:
fifty tabs,
little understanding.
Too much exploitation creates:
one beautifully understood source,
possibly the wrong source.
Strong foraging alternates.
Explore.
Find a promising patch.
Exploit.
Notice a gap.
Explore again.
Then deepen.
The practical skill is simple.
Before another search move, ask:
Do I need another possibility, or do I need more depth on the best possibility I already have?
Those are different jobs.
If you still do not know what kinds of explanation exist:
explore.
If you have identified the right mechanism but do not understand it:
exploit.
If five sources repeat the same overview:
stop broad reading and go deeper.
If one source is carrying the entire conclusion:
explore again.
This is information strategy becoming deliberate.
Worth learning because: effective search depends on knowing when another new source is more valuable than deeper understanding of the strongest source already found.
6. Learn to Search for the Information That Reduces the Most Uncertainty
Some information is more valuable than other information.
Imagine three possible explanations for a student’s weak Mathematics performance:
A: content knowledge is weak.
B: method selection is weak.
C: time pressure is causing execution failure.
You could search broadly for:
“How to improve Mathematics.”
Low information gain.
Or test:
Can the student solve direct, untimed questions accurately?
If yes, Explanation A weakens immediately.
One observation removes a large part of the hypothesis space.
That is an efficient information move.
Developmental information-search research often studies this principle using tasks such as Twenty Questions. Constraint-seeking questions can eliminate classes of possibilities rather than checking one hypothesis at a time, and Török and colleagues argue that search efficiency should be distinguished from mere success.
This gives students an important question:
What could I learn next that would change my search most?
Not:
What else can I read?
Suppose you are researching a current policy.
You do not know whether it applies in Singapore.
Before reading twenty commentaries, find the jurisdiction.
One fact can collapse most of the search space.
Suppose a Science student cannot interpret an experiment.
They do not know which variable changed.
Resolve that first.
Suppose a GP student is evaluating a dramatic statistic.
Find the denominator.
Perhaps the entire argumentative value changes.
High-value search is often less about gathering more and more about locating the piece that narrows the remaining possibilities fastest.
Worth learning because: the best next search move is often the one that eliminates the largest number of plausible interpretations rather than the one that retrieves the largest amount of information.
7. Learn to Preserve Provenance While You Search
Students lose sources constantly.
They remember:
“I read somewhere that…”
That phrase should make a serious learner nervous.
Where?
Original study?
Blog?
AI summary?
News report?
Teacher?
Textbook?
Government document?
If the claim matters, the route matters.
Information Foraging therefore needs a provenance habit.
Not a full academic citation system for every ten-year-old.
Just:
Where did this come from?
At Primary level, that may mean keeping the book title or webpage.
At Secondary level:
author,
organisation,
page title,
date.
At JC level:
DOI,
paper,
publisher,
dataset,
retrieval route.
When using AI, preserve something even more important:
Did the information come from the AI’s generated prose, or from a source the AI led you to?
Those are not equivalent.
A search trail also helps when the learner gets lost.
Search A produced terminology.
Search B found a review.
The review led to Study C.
Study C cited Dataset D.
Now the reasoning can be reconstructed.
Without provenance, the student possesses orphan facts.
They may be correct.
They cannot easily be audited.
This is a clean boundary with Verification.
Verification determines whether the source or claim deserves trust.
Information Foraging keeps the source attached long enough for Verification to perform that job.
Worth learning because: information that has lost its provenance becomes much harder to verify, compare, revisit or responsibly use.
8. Learn to Match the Source Type to the Information Job
Students sometimes ask one source to do everything.
A Wikipedia page to establish a legal requirement.
A news report to establish the exact research method.
One small experiment to describe the whole literature.
A systematic review to explain a Primary-school concept.
Wrong source job.
Different information needs call for different patches.
Need an official rule?
Go to the responsible authority.
Need to know exactly what researchers did?
Find the primary paper.
Need the broad state of a mature literature?
Look for a systematic review or authoritative synthesis.
Need a definition?
A suitable textbook, reference work or official glossary may be enough.
Need a current event?
Recent reporting matters.
Need the raw numbers?
Find the dataset or primary report.
Need to understand a concept for the first time?
A high-quality explainer may be pedagogically superior to the original research paper.
This is one of the quieter marks of a strong researcher.
They do not ask:
Which source is best?
They ask:
Best for which job?
A primary paper can be excellent for method detail and poor for estimating the entire evidence base.
A systematic review can be excellent for synthesis and too high-level to reconstruct one experiment.
A government page can be authoritative about its own rules and not authoritative about an unrelated scientific mechanism.
Source quality is partly relational.
Fit the source to the question.
This also helps prevent false sophistication.
Students sometimes believe the most technical source is always the best.
No.
If the job is to understand what osmosis means at Secondary level, a clear curriculum-aligned explanation may be much more useful than an advanced membrane-transport paper.
The best patch is the patch suited to the current information job.
Worth learning because: search quality improves when learners select sources according to the kind of evidence, explanation or authority the question actually requires.
9. Learn When to Switch Patches—and When to Stop Searching
Search can become addictive.
One more article.
One more prompt.
One more result page.
Perhaps the next source will finally remove all uncertainty.
Usually not.
At some point, the marginal value falls.
Pirolli’s information-foraging work explicitly models this leave-or-stay decision as a comparison between the expected benefit of continuing in the current information patch and going elsewhere.
Students need a version they can use.
Stay when:
the source is still answering important unresolved parts of the question.
Switch when:
new material repeats what you already know,
the source moves away from your information gap,
the source lacks the detail required,
the terminology shows you are in the wrong patch,
or another source type would answer the remaining question better.
Stop when:
additional searching is unlikely to change the answer, decision or confidence enough to justify the extra time.
That last sentence matters.
Search does not need to eliminate all uncertainty.
Sometimes:
three independent high-quality sources agree,
the mechanism is understood,
the important counterposition has been checked,
and additional results are repeating the same evidence.
Stop.
Other times:
sources conflict on the exact point that determines the decision.
Do not stop merely because the deadline is near.
Reframe.
What distinguishes the conflicting sources?
Population?
Date?
Definition?
Method?
Now a new information gap appears.
Stopping is therefore not exhaustion.
It is a decision about marginal value.
Worth learning because: knowing when another source is unlikely to change the result protects learners from both premature closure and endless low-yield browsing.
10. Learn to Return From Search With a Changed Model, Not a Pile of Tabs
Search has a purpose.
Something should change.
Before:
I did not know X.
After:
I know Y.
Or:
I thought A.
Now the evidence favours B.
Or:
I still cannot decide because C remains unresolved.
Or:
the question was too broad and must be reframed.
That is a return.
Twenty saved links are not a return.
A long AI transcript is not a return.
A page of copied quotations is not a return.
Information must re-enter the learner.
A powerful closing habit is:
What changed because I searched?
Then:
What remains uncertain?
Then:
What is the next action?
This is the handoff to Top 10 Synthesis Skills Worth Learning.
Information Foraging finds and triages.
Synthesis integrates.
Verification checks.
Studying turns the result into durable capability.
Problem Framing may receive a new question back.
The search loop therefore becomes:
GAP → SEARCH → RETURN → UPDATED GAP
Search is not a detour from thinking.
Done properly, it is part of thinking.
Worth learning because: information has educational value only when the search changes the learner’s understanding, decision or next move rather than merely increasing the number of materials collected.
The Top 10 Information Foraging Skills as One System
The Wintour House route is:
GAP → QUERY → SCENT → LANDSCAPE → EXPLORE/EXPLOIT → INFORMATION GAIN → PROVENANCE → SOURCE-TYPE FIT → SWITCH/STOP → LEARNING RETURN
The quieter version is:
Know what you are missing. Search in the language of the thing you need. Use clues to decide where attention is worth spending without confusing relevance with truth. Look around before committing. Go deep when the patch is rich and move when it is not. Search first for information that removes the most uncertainty. Keep the source attached. Use the right kind of source for the job. Stop when more searching stops changing the answer. Then return with a better model, not more tabs.
That is information foraging.
Not Googling quickly.
Not opening everything.
Not trusting the top result.
Not collecting PDFs.
Not asking AI until it says something agreeable.
Not browsing until time runs out.
Information foraging is search effort under control.
Information Foraging Is Not the Same as Questioning
Top 10 Questioning Skills Worth Learning owns the construction of useful questions.
Information Foraging begins when the learner asks:
Where and how should I search for the information that could answer this question?
Questioning creates the request.
Foraging routes the request through an information environment.
Information Foraging Is Not the Same as Problem Framing
Top 10 Problem-Framing Skills Worth Learning asks:
What problem are we actually dealing with?
Information Foraging asks:
Which unresolved parts of that problem require external information, and where is the highest-value place to look next?
Problem Framing owns the problem representation.
Foraging owns the search trajectory.
Information Foraging Is Not the Same as Verification
Top 10 Verification Skills Worth Learning asks:
Should this claim be accepted?
Information Foraging asks earlier:
Which claims and sources deserve scarce inspection time?
Strong information scent can lead to a source that later fails verification.
That is not a contradiction.
Relevance got us to the door.
Verification decides whether we enter.
Information Foraging Is Not the Same as Critical Thinking
How to Improve Critical Thinking owns broad evaluation of claims, evidence, assumptions, alternatives and inference.
Information Foraging makes sure the learner does not conduct all that beautiful thinking inside one arbitrarily selected information patch.
Critical Thinking evaluates what is found.
Foraging governs where to look.
Information Foraging Is Not the Same as Synthesis
Top 10 Synthesis Skills Worth Learning owns integration across distributed information.
Foraging comes before and during that work.
Which sources should enter the candidate pool?
Where should the learner search next because the synthesis contains a gap?
Synthesis constructs the larger knowledge object.
Foraging supplies and replenishes its information corridors.
Information Foraging Is Not the Same as Research Skills
Research is broader.
Question formulation.
Methods.
Ethics.
Evidence evaluation.
Analysis.
Citation.
Writing.
Replication.
Information Foraging is one reusable engine inside research:
how scarce search attention is allocated across an information environment.
That narrow ownership protects a future Research Skills page from cannibalisation.
Information Foraging Is Not the Same as Browsing
Browsing can be exploratory.
Curiosity has value.
Serendipity has value.
Not every information encounter needs an explicit target.
But when a learner has a defined job, purposeful foraging differs from wandering.
The question becomes:
What is the expected value of the next click for the thing I am trying to understand or decide?
That makes search accountable.
For Primary Students
Primary Information Foraging should not begin with Boolean operators.
It begins with purposeful looking.
A child asks:
“Why do penguins not fly?”
Good.
Now:
Where could we look?
Science book?
Index?
Animal encyclopedia?
Teacher?
A suitable educational website?
Which title looks promising?
Which heading?
What keyword?
If the first page only describes where penguins live, should we keep reading?
Probably not.
Move.
Children can also learn the difference between a broad search and a narrowing search.
“Animals.”
Huge.
“Penguin wings swimming flight.”
Much better.
Developmental information-search research gives educators an encouraging but careful message. Török and colleagues note that findings about children’s search competence vary greatly with task design and that young children can show adaptive information-seeking when tasks are developmentally suitable. They argue that search quality should be understood through both efficiency and eventual effectiveness rather than judging children solely by whether they happened to reach the answer.
The Primary goal is simple:
Know what you are looking for. Choose a useful clue. Check whether the place is helping. Move if it is not. Remember where the information came from.
That is enough to build the architecture.
For Secondary Students
Secondary learners can become deliberate query engineers.
Concept.
Relationship.
Condition.
Synonym.
Source type.
Date.
They can learn:
broad pass,
then narrow pass.
They can compare:
official source,
educational explanation,
research article,
news report.
They should also begin keeping a search trail.
Question.
Queries tried.
Useful sources.
Rejected routes.
Remaining gap.
Not because every homework task needs a laboratory notebook.
Because longer investigations become difficult to control otherwise.
Secondary learners should also become aware of copy cascades.
Five sites may repeat the same claim.
That is not five independent sources.
Find the origin.
This is not yet the final Verification judgement.
It is basic information-landscape literacy.
For JC Students
JC students should treat information search as part of argument quality.
A GP essay about technology and employment cannot rely on:
“AI jobs statistics.”
The learner needs corridors.
Current labour data.
Research on task exposure.
Historical analogues.
Firm adoption.
Wage effects.
Jurisdiction.
Counterevidence.
Then:
Which question actually matters to the thesis?
Search should narrow toward that.
JC learners can also begin citation chaining.
Find a strong review.
Inspect the studies it depends on.
Find a key primary paper.
Look forward to later work that cites it.
Now the learner is moving through an evidence network rather than relying entirely on one search engine’s ranking.
The maturity signal is not:
“I found many sources.”
It is:
“I know why these are the sources I chose to inspect.”
Information Foraging in Mathematics
Mathematics search often fails because students search the surface wording of a problem.
“A shop sells three bags…”
They want the exact worked answer.
A stronger learner searches by structure.
Simultaneous equations.
Constant rate.
Similar triangles.
Invariant.
Ratio change.
Conditional probability.
Search for the mathematical object.
Not the story decoration.
This is where Pattern Recognition and Abstraction help.
Information Foraging uses their output to locate relevant resources.
A useful Mathematics search should eventually make the learner less dependent on searching the exact problem.
If every new surface story triggers another Google query, transfer has not occurred.
Information Foraging in Science
Science search needs source-role awareness.
Need a curriculum explanation?
Use an educational source suited to the level.
Need current scientific evidence?
Look for the research literature.
Need a safety or regulatory rule?
Use the responsible authority.
Need the exact experimental method?
Primary source.
Need the broad state of evidence?
Review.
This protects students from one common mistake:
reading a high-level public explainer and then treating it as though it contains the evidential resolution of a research paper.
Or the reverse:
opening an advanced research article when the actual need is simply understanding a Primary Science concept.
Search resolution should fit the learning job.
Information Foraging in English and GP
English and GP students often need external evidence.
Statistics.
Examples.
Cases.
Definitions.
Context.
The danger is argument-led search.
Student decides:
“My conclusion is true.”
Then searches only for support.
That is not foraging.
That is evidence shopping.
A better route is:
What evidence would support this?
What evidence would weaken it?
Which population matters?
Which source would actually know?
Does the opposing case have a credible patch?
Search can therefore function as intellectual resistance.
It should expose the thesis to a wider landscape than the learner’s first preferred story.
Information Foraging in Studying
Studying can become a search problem.
The student does not understand one concept.
They open the internet.
Thirty minutes later:
another concept,
another video,
another tab.
The original gap remains.
A strong study-foraging loop is:
What exact thing is blocking me?
Search only for that.
Once resolved:
close the search environment.
Return to the task.
This protects the live Top 10 Studying Skills Worth Learning owner.
Studying controls the learning programme.
Information Foraging is a support action used when external information is required.
Search should repair the study route.
Not replace studying with browsing.
Information Foraging in Research
Research makes search discipline visible.
A serious search can record:
database,
query,
date,
filters,
inclusion criteria,
excluded source types,
citation chains.
That level of formality is not needed for ordinary homework.
But the underlying habits scale down beautifully.
What did I search?
Why did I search it?
What did it reveal?
What new term did I discover?
Which source does everyone else cite?
Where does disagreement begin?
When did additional searching stop changing the evidence picture?
A systematic review and meta-analysis by Boetje and colleagues, first published online in 2025 and appearing in the Review of Educational Research, synthesised 69 controlled pretest–posttest publications from higher education and identified seven interrelated instructional design principles: learning task, instruction, modelling, practice, learning activities, support and feedback. The evidence base is concentrated in undergraduate settings, so it should not be projected directly onto Primary or Secondary learners, but it strongly supports treating information problem solving as teachable rather than assuming access to digital tools creates competence automatically.
That is a useful boundary.
Access is not competence.
Information Foraging in the Age of AI
AI changes the search landscape profoundly.
Traditional search often says:
Here are ten doors.
AI often says:
Here is the answer.
That is convenient.
It also compresses several information-foraging decisions into one invisible process.
Which sources were considered?
Which patch dominated?
Which interpretation was discarded?
Was current information used?
Was the answer synthesised from independent evidence or repeated claims?
The learner sees less of the landscape.
This can reduce friction.
It can also reduce useful intellectual work.
A randomized study involving 91 university students compared ChatGPT 3.5 with Google during research on nanoparticles in sunscreen. The LLM group reported lower cognitive load but produced weaker reasoning in their final analyses. The study is one task and one student population, so it should not be treated as a universal verdict on AI search; it is nevertheless a useful warning that making information gathering easier does not automatically make the resulting reasoning deeper. See Stadler, Bannert and Sailer, “Cognitive ease at a cost”.
So the Wintour rule is not:
Use Google, not AI.
Nor:
Use AI, not Google.
It is:
Use the tool that gives the best next information move—and keep enough control of the search landscape to know what the tool has hidden.
AI can be an excellent scout.
Ask:
“What terminology should I search for?”
“What are the major competing explanations?”
“What source types would answer each part?”
“What is the original study behind this claim?”
“Give me three independent evidence corridors rather than one synthesis.”
“What would I search outside this conversation to verify the central claim?”
Then leave the AI patch when appropriate.
Go to the source.
Come back if useful.
The learner should not be trapped inside one conversational patch simply because it is comfortable.
The Information Foraging Paradox: Faster Search Can Produce Slower Learning
AI gives the answer instantly.
Excellent.
But perhaps the learner never:
compared alternatives,
noticed the vocabulary,
saw the evidence structure,
or learned where information lives.
The search was faster.
The next independent search may not be.
Efficiency of retrieval and development of search competence are different outcomes.
The Information Foraging Paradox: More Sources Can Produce Less Knowledge
Twenty tabs.
Ten papers.
Seven summaries.
The learner’s working model becomes noisier.
Source count is not a measure of understanding.
After a point, new material can increase duplication and contradiction faster than it increases useful information.
Search quality needs compression.
What has actually changed?
The Information Foraging Paradox: The Most Relevant Result Can Be the Most Dangerous
A result matches your exact phrasing.
Wonderful information scent.
Perhaps because the page was written to capture precisely that search phrase.
Relevance increased.
Credibility did not automatically increase.
Strong scent deserves attention.
Not submission.
The Information Foraging Paradox: A Good Search May Look Inefficient Before It Becomes Efficient
Weak learner:
clicks first answer instantly.
Fast.
Strong learner:
defines the gap,
scans several patches,
identifies terminology,
then commits.
Slower beginning.
Much faster downstream.
They avoid:
wrong rabbit holes,
duplicate sources,
rewriting the question halfway through,
and discovering at the end that the evidence answered something else.
Good search often spends a little more attention near the entrance so that less is wasted deep inside the wrong corridor.
The Information Foraging Paradox: Knowing When to Stop Is Part of Knowing How to Search
People praise curiosity.
Correctly.
But professional search includes stopping.
A physician cannot read every paper before acting.
A student cannot open every page before submitting.
A researcher eventually closes the literature search under defined rules.
A learner needs a decision rule.
Not certainty.
Enough information for the current job.
Then return to the world.
The Wintour House Test: Does Information Foraging Survive When AI Can Retrieve Anything Instantly?
Suppose AI becomes extraordinary.
Every paper indexed.
Every webpage read.
Every book searchable.
Every relevant paragraph retrievable.
Every language translated.
Every answer generated instantly.
Does Information Foraging disappear?
No.
Because somebody still has to decide:
what is actually unknown,
which wording exposes the right information landscape,
which apparent scent deserves attention,
whether the search is still too narrow,
when breadth should become depth,
which missing fact would reduce the most uncertainty,
which source type owns the relevant authority,
whether independent information corridors exist,
when another search is worth its cost,
and what changed after the information returned.
Retrieval can become nearly free.
Attention cannot.
That is why Information Foraging belongs permanently in the Skills Worth Learning series.
The mature learner can eventually say:
I know what information I am missing. I can translate that gap into useful search moves. I can use information scent without confusing relevance with credibility. I explore enough of the landscape before committing. I know when to broaden and when to go deep. I search first for information that removes the most uncertainty. I preserve provenance, match source type to information job, and recognise diminishing returns. Most importantly, I return from search with a changed understanding rather than a larger collection of tabs.
That is Information Foraging becoming intellectual navigation.
Research Anchors
The ten skills above are a Wintour House editorial synthesis, not a claim that information-science or cognitive-development research has validated one universal ten-part taxonomy of Information Foraging.
Peter Pirolli’s 2005 Cognitive Science paper remains a foundational anchor. Building on information-foraging theory, it models web navigation through information scent—proximal cues linked to users’ information goals—and models site-leaving decisions as continuing assessments of the expected benefit of remaining in a patch versus moving elsewhere.
Török, Stanciu and Ruggeri’s 2024 Child Development Perspectives paper supplies an important developmental correction. It argues that information-search competence should distinguish efficiency, effectiveness and success: a successful outcome can occur through luck, while an informative search strategy may occasionally fail because of environmental contingencies. The paper also discusses constraint-seeking questions and expected information gain as ways of thinking about how effectively a search move reduces uncertainty.
Boetje and colleagues’ systematic review and meta-analysis, first published online in 2025, synthesised 69 peer-reviewed publications from 2000–2023 with controlled pretest–posttest designs in higher education. The authors identified seven instructional design principles—learning task, instruction, modelling, practice, learning activities, support and feedback—and note that empirical work has disproportionately concentrated on information search and selection. Because the corpus concerns undergraduate education, it supports the teachability of complex search competence more strongly than it supports any precise Primary or Secondary intervention claim.
AI-era research adds useful but context-specific boundaries. Stadler, Bannert and Sailer’s study randomly assigned 91 university students to use ChatGPT 3.5 or Google while researching nanoparticles in sunscreen. LLM users reported lower cognitive load but showed weaker reasoning in their analyses. That does not establish that AI search generally harms learning; it demonstrates that easier retrieval and deeper inquiry are separable outcomes.
The strongest defensible Wintour House conclusion is therefore:
Information Foraging is not merely finding information. It is disciplined allocation of search attention: define the information gap, formulate informative search moves, use relevance cues without confusing them with credibility, explore the landscape before committing, balance breadth and depth, seek high-information-gain evidence, preserve provenance, fit source type to information need, recognise diminishing returns and close the search loop by returning with a better understanding, decision or next question.
