Why Kids Can't Apply What They Know to New Situations (and How to Fix It)

A parent recently described this to me: her 11-year-old had just received a 94% on a math test covering percentages. A week later, at dinner, she asked him to figure out a 20% tip on a $35 bill. He stared at the table. He had no idea where to start.

This is not a memory problem. The knowledge is there. It’s a transfer problem — and it’s one of the most documented, most underappreciated challenges in how children learn. The research goes back over a century, and the solutions are specific, teachable, and almost never used in standard instruction.

Key Takeaways

• Transfer of learning — applying knowledge to new situations — is far harder than acquiring the knowledge in the first place, and it doesn’t happen automatically.
• Edward Thorndike’s foundational research showed that only skills with shared “identical elements” transfer reliably; skills that look similar but differ structurally often don’t.
• The “inert knowledge” problem, named by philosopher Alfred North Whitehead in 1929, describes knowledge that exists in memory but fails to activate when needed in real-world contexts.
• Gick and Holyoak’s landmark 1983 research showed that giving children a single analogy rarely produces transfer — but giving two or more analogies, with discussion, dramatically increases it.
• Interleaved practice (mixing problem types within a study session) produces significantly better transfer than blocked practice (mastering one type before moving on), even when it feels harder in the moment.

What Thorndike Got Right — and What He Missed

In 1901, Edward Thorndike and Robert Woodworth published a series of experiments that upended the prevailing theory of education. The assumption at the time was that studying subjects like Latin and geometry improved general mental discipline — that they trained the mind in a way that would benefit any learning task. Thorndike and Woodworth tested this and found it wasn’t true.

Transfer, they showed, occurs only when two tasks share identical elements — specific stimulus-response connections in common. Training in estimating the area of rectangles improved estimation of other rectangles but not triangles. Estimating length in inches didn’t improve estimation in centimeters. The overlap had to be structural, not merely superficial.

This finding was sobering in 1901 and remains relevant now. Most school instruction assumes that once a child learns a concept in one context, they can readily apply it in others. Decades of research suggest this assumption is frequently wrong — and that the gap between what children know and what they can use is much wider than test scores suggest.

The research on retrieval practice and the testing effect shows that frequent low-stakes retrieval strengthens memory consolidation — but even well-consolidated knowledge can remain context-locked. Retrieval and transfer are related but distinct skills.

Near Transfer vs. Far Transfer: A Critical Distinction

Transfer researchers draw a line between two very different phenomena.

Near transfer is applying a skill in a context that closely resembles the original learning context. A child who learns to add fractions with the same denominator and then successfully adds fractions with different denominators is demonstrating near transfer. The surface features and the underlying procedure are similar enough that the skill generalizes with relatively little additional work.

Far transfer is applying knowledge in a context that looks quite different from the original. The restaurant tip scenario is far transfer: the surface features (eating out, real money, social context) look nothing like the classroom percentage-calculation scenario (printed worksheet, abstract numbers, test conditions).

Far transfer is rare. Rarer than most educators expect, and far rarer than most curricula assume. John Bransford, Ann Brown, and Rodney Cocking documented this comprehensively in the 2000 National Research Council report How People Learn (https://www.nap.edu/catalog/9853/how-people-learn-brain-mind-experience-and-school-expanded-edition), which synthesized decades of learning science. The report’s central finding: students who can perform well on assessments frequently cannot apply the same knowledge to novel problems, because they have learned the procedure without learning the underlying principle.

Transfer Type	Description	Example	Frequency in Real Learning
Near transfer	Same procedure, slightly different surface	Fraction addition → different denominators	Common with explicit review
Intermediate transfer	Shared principle, different content domain	Percentage in math → percentage in science data	Requires explicit instruction
Far transfer	Abstract principle applied across very different domains	Ratio reasoning → cooking → map reading → probability	Rare without deliberate teaching
Zero transfer	No generalization from original learning	Long division → tip calculation (common)	Disturbingly frequent
Negative transfer	Prior learning actively interferes	English spelling rules → French spelling	Common across languages

Source: Bransford et al. (2000), How People Learn; Thorndike & Woodworth (1901), Psychological Review.

The Inert Knowledge Problem

Alfred North Whitehead coined the phrase “inert knowledge” in his 1929 essay collection The Aims of Education. He was describing something he found alarming in British schooling: students who possessed facts and procedures that never became active — knowledge that sat in memory like furniture in a locked room, present but inaccessible when needed.

John Bransford and colleagues gave this concept empirical teeth over decades of research at Vanderbilt University (https://peabody.vanderbilt.edu/). Their studies showed that students who received traditional instruction on a topic could answer specific questions about it but failed to spontaneously use that knowledge when facing problems where it was relevant. The knowledge existed; it just didn’t activate.

This is not about motivation or effort. Students who are highly motivated still show the inert knowledge effect. It’s a problem of how knowledge is encoded. When a child learns that 20% equals one-fifth in the specific context of a math worksheet, the knowledge gets stored with contextual cues tied to that worksheet: the font, the instructions, the classroom setting. When the context shifts radically — a restaurant, real money, social pressure — those retrieval cues are absent, and the knowledge doesn’t surface.

Why Analogies Need to Come in Pairs

Keith Holyoak and Mary Gick ran a series of experiments in 1980 and 1983 that are now considered foundational in the transfer literature. They presented college students with Duncker’s radiation problem — a classic puzzle about how to destroy a tumor without damaging surrounding tissue using radiation. The correct solution involves using multiple low-intensity rays from different angles to converge at the tumor.

Gick and Holyoak found that presenting students with one analogous story (a military story where a general captures a fortress by sending small groups of soldiers down multiple roads to converge at the center) produced spontaneous transfer to the radiation problem in only about 20–30% of cases. Presenting two analogous stories, followed by a prompt to compare them, raised spontaneous transfer to over 60%. The key was that comparing two analogies forced students to extract the abstract principle (convergence of weak forces) rather than storing the specific story.

This finding has been replicated with children across a range of ages. Usha Goswami’s research on analogical reasoning in children (summarized in her 2001 Psychological Bulletin review) shows that even young children can reason analogically when the relational structure is made explicit — but they default to surface similarity when left to their own devices.

Dedre Gentner’s work on analogical learning (https://psychology.northwestern.edu/people/faculty/core/profiles/dedre-gentner.html) has further shown that comparison — explicitly asking “how is this like that?” — is among the most powerful tools for promoting abstract principle extraction in learners of all ages.

The implication for parents is direct: when teaching a concept, don’t just show one example. Show two or three examples from different domains and ask “what do these have in common?”

Why Procedural Knowledge Fails to Transfer

There is a specific pattern worth naming: children who learn a procedure — the steps to do long division, the formula for finding area, the sequence for writing a paragraph — often learn it as a script. They know the moves, in order, in the context where they were taught.

Procedural knowledge is highly contextual. It’s efficient for familiar problems but brittle under variation. A child who has learned to find the area of a rectangle using the formula length × width may be completely stumped when asked to find the area of an irregular shape that requires decomposition — even though the same multiplication is involved.

This brittleness happens because the child never learned why the formula works — what relationship it is capturing. The principle (area = number of unit squares that fit inside the shape) was never extracted from the procedure (multiply the two numbers). Without the principle, there is no handle for transfer.

Research on metacognition and self-regulated learning in children shows that students who monitor their own understanding — who notice when they know how but not why — are significantly better at initiating transfer in novel situations. Metacognitive awareness acts as the trigger that says “I’ve seen something like this before.”

Interleaved Practice Produces More Transfer

One of the most consistent findings in the transfer literature is the superiority of interleaved practice over blocked practice for promoting transfer.

Blocked practice: complete all problems of type A, then all problems of type B, then all problems of type C. This is the standard approach in most textbooks and most classroom instruction.

Interleaved practice: mix problem types A, B, and C within the same practice session, in random or varied order.

Nate Kornell and Robert Bjork’s 2008 study in Psychological Science showed that interleaved practice consistently outperformed blocked practice on transfer tests, even though students typically felt they were learning less during interleaved sessions. (The sensation of difficulty is not a reliable indicator of learning.) The effect held across age groups and content domains.

The mechanism appears to be that interleaved practice forces learners to identify which strategy applies before executing it — a discrimination task that is absent in blocked practice. In blocked practice, the learner knows which procedure is relevant before reading the problem. In interleaved practice, they must retrieve the procedure from among competitors. This retrieval with discrimination is closer to real-world problem solving, where the relevant knowledge isn’t labeled in advance.

Strategies That Actually Produce Transfer

Use multiple worked examples before practice

Before asking a child to practice a type of problem, show at least two worked examples from different contexts. Not two fraction problems — one fraction problem and one involving percentages of a recipe. The surface difference forces attention to the shared structure.

Ask “where else could this apply?” after every lesson

This one question, asked consistently, appears to strengthen the formation of abstract schemata. It prompts the child to actively search for structural matches rather than passively encoding the specific example. It doesn’t have to take long — three minutes at the end of a homework session is enough.

Teach the principle before the procedure

Before showing the steps, explain the why. Before long division, make sure the child can explain what division is capturing (dividing a group into equal-sized parts). Before percentage calculations, make sure they can explain what a percentage represents (parts per hundred). This gives the procedure something to hang on.

Use comparison tasks explicitly

Take two problems that look different on the surface but share the same underlying structure. Ask the child to explain how they’re alike. This is the technique that drove Gick and Holyoak’s transfer gains from 30% to 60%. It works with children as young as 7 when the comparison is kept concrete.

Practice in varied contexts deliberately

A child who learns something only at the kitchen table may retrieve it only in kitchen-table contexts. Practice the same concept across locations, formats, and applications. The restaurant tip. The science class data set. The card game odds. The same underlying ratio-reasoning skill — practiced in genuinely different contexts — becomes more flexibly accessible.

What to Watch For Over the Next 3 Months

Month 1: After introducing a new concept in one context, try presenting a genuinely different application within the same week. Not the same homework problem with different numbers — a different context entirely (cooking, sports, a board game). Note whether the child recognizes the connection or needs prompting. Most will need prompting at first.

Month 2: Introduce intentional comparison tasks. After a math problem and a science activity that use the same reasoning, ask: “Did you notice these were similar? What did they have in common?” At first, children typically don’t notice. By the second month of consistent practice, some will start noticing spontaneously.

Month 3: If transfer is improving, you’ll see the child occasionally ask “Is this like that thing we did with the recipes?” without prompting. That spontaneous analogical recognition is what successful transfer education looks like. If you’re still not seeing any transfer after three months of varied practice, the underlying concept may not be fully understood — consider stepping back to the concrete “why” before the procedural “how.”

Red flag: A child who can recite a procedure perfectly but draws a complete blank when the context shifts slightly (not radically — slightly) is showing extreme procedural encoding. This warrants a conversation about understanding, not a push for more practice.

Frequently Asked Questions

My daughter memorizes her math homework perfectly but fails the test. Is this a transfer problem?

Possibly. If her homework is blocked practice on one problem type, and the test mixes problem types, she may not be recognizing which strategy applies — a discrimination skill that blocked practice doesn’t develop. Try interleaving her practice at home: mix problem types from different recent topics in each session, even briefly.

How young can kids start doing comparison tasks to build transfer?

Younger than you’d think. Goswami’s research shows children as young as 4 can reason analogically when relationships are made explicit and concrete. For 5–7-year-olds, comparison tasks work best with physical objects or vivid stories. Abstract comparison (“both use multiplication”) becomes accessible around ages 9–11 for most children.

Is far transfer something schools should be teaching more explicitly?

Yes — and there’s reasonable consensus on this in learning science, even if curricula lag behind. The National Research Council’s How People Learn report, the OECD’s work on transferable competencies (https://www.oecd.org/education/), and decades of cognitive psychology research all point to the same conclusion: transfer should be an explicit instructional goal, not an assumed byproduct of content mastery.

What subjects show the best evidence for transfer when taught well?

Mathematics and reading comprehension have the strongest transfer research bases. Math reasoning skills — particularly ratio, proportion, and algebraic thinking — show robust far transfer when taught with principled understanding. Reading comprehension strategies (identifying main idea, drawing inferences) transfer across content areas when practiced in varied texts. Science reasoning skills show promising but more limited transfer evidence.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Thorndike, E. L., & Woodworth, R. S. (1901). “The influence of improvement in one mental function upon the efficiency of other functions.” Psychological Review, 8(3), 247–261. https://doi.org/10.1037/h0074898
2. Gick, M. L., & Holyoak, K. J. (1983). “Schema induction and analogical transfer.” Cognitive Psychology, 15(1), 1–38. https://doi.org/10.1016/0010-0285(83)90002-6
3. Bransford, J. D., Brown, A. L., & Cocking, R. R. (Eds.). (2000). How People Learn: Brain, Mind, Experience, and School (Expanded ed.). National Academy Press. https://www.nap.edu/catalog/9853
4. Kornell, N., & Bjork, R. A. (2008). “Learning concepts and categories: Is spacing the ‘enemy of induction’?” Psychological Science, 19(6), 585–592. https://doi.org/10.1111/j.1467-9280.2008.02127.x
5. Goswami, U. (2001). “Analogical reasoning in children.” Psychological Bulletin, 127(2), 169–196. https://doi.org/10.1037/0033-2909.127.2.169
6. Whitehead, A. N. (1929). The Aims of Education and Other Essays. Macmillan.
7. OECD. (2019). OECD Learning Compass 2030: A Series of Concept Notes. OECD Publishing. https://www.oecd.org/education/2030-project/