Concept/Memory and Learning/No. 0561

Learning Transfer

Learning transfer is the use of knowledge or skill in a setting beyond where it was learned. Studied in psychology by Edward Thorndike and Robert Woodworth, it depends on shared task features, cues that bring learning to mind, and recognizing when old knowledge applies.

Also called Transfer of Learning

a concept: name it

Evidence: Well established
Read: 6 min
Links: 11 connections
Useful when: Designing products · Judging people and hiring · Learning and memory · Organizations and bureaucracy

01You've seen this when…

in life
You can calculate percentages on a worksheet. At a store, a 20% discount followed by another 10% off leaves you reaching for your phone.
at work
The onboarding quiz goes smoothly. Two days later, a customer describes the same issue in different words, and you miss the fix.
out in the world
Volunteers practice evacuation procedures in a community center. At a street festival, they face a blocked exit and have to adapt the same rules to an unfamiliar layout.

02The idea

A lesson can be easy to recall in the classroom and hard to use anywhere else. The worksheet supplies a topic, familiar wording, and a clear instruction. Everyday problems arrive with those cues missing. Part of the job is figuring out which lesson applies.

Learning transfer is the carryover from earlier learning to another task or setting. It can happen almost automatically: learning one word processor makes another easier to navigate. It can also require deliberate work, such as recognizing that a dispute over household chores has the same incentive problem as a team project.

Researchers distinguish near and far transfer. Near transfer involves closely related tasks. Far transfer crosses larger differences in subject, setting, or the kind of performance required. Distance has several dimensions: a task might use the same calculation while changing the wording, workplace, and tools.

Three conditions help explain the carryover. The new task shares something useful with the old one. The person retrieves the relevant knowledge. They recognize how to adapt it. Shared structure can remain hidden beneath different surface details, which makes analogical reasoning an important route to transfer.

A learning curve tracks improvement with practice. Transfer adds another question: how much of that improvement survives a change of task or setting?

03Why it matters

Training usually aims at performance somewhere else. A safety course is supposed to help during an emergency. A statistics class is supposed to help someone interpret an unfamiliar claim. A management workshop is supposed to change what happens in a difficult conversation.

That makes transfer a practical test of whether instruction is doing its job. A high score on familiar exercises can coexist with weak performance on new cases. The exercises may supply clues that the eventual situation leaves out.

Transfer also changes how people assess expertise. Someone who handles the standard case smoothly may need support when the details change. Testing a few unfamiliar cases reveals the reach of their knowledge and helps identify the right next step: more foundational practice, better recognition cues, or experience adapting the skill.

04A worked example

In 1980, psychologists Mary Gick and Keith Holyoak studied transfer using two problems with very different settings. One described a general trying to capture a fortress. The approach roads were mined, and each could safely carry only a small group. The solution was to split the army and have the groups converge on the fortress.

Participants then faced a medical problem: a strong ray could destroy a tumor but would damage the healthy tissue along its path. Several weaker rays, aimed from different directions and meeting at the tumor, offered a solution.

What it looks like A military story and a medical puzzle seem to require separate bodies of knowledge. Remembering the general’s plan gives no obvious instruction for treating a tumor.

What’s actually going on Both problems share a structure: divide a force into smaller parts, send those parts along different paths, and concentrate their effect at the destination. In a comparison from the experiments, the solution rate was 30% when the story was supplied without a hint and 75% when participants were prompted to use it. Many participants needed help recognizing the connection.

What would have helped Making the shared structure explicit, then asking participants to map each part onto the new problem. The roads correspond to paths through tissue; the small groups correspond to weaker rays. In teaching, comparing cases this way can give learners a pattern they can search for later.

05Where people trip up

Practice can become tied to one format. Ten exercises with identical wording teach the subject and the worksheet’s cues together. Change the wording, representation, or setting once the basic skill is secure. A percentage problem might appear as a receipt, a chart, or a short email. These changes help reveal which parts of the skill survive.
Knowing a method leaves recognition work unfinished. A learner may understand a procedure and still miss the moment to use it. Ask them to identify the features that make the method relevant, then compare a case where it applies with one where it fails. A useful explanation names both the pattern and its boundaries.
Too much novelty can overwhelm a beginner. Transfer requires something solid to carry over. An unfamiliar task with several new demands can consume working memory before the learner reaches the relevant step. Begin with a worked example, then gradually remove guidance and change the context.
Broad claims can outrun narrow evidence. Faster performance on a practiced puzzle supports a claim about that puzzle. A claim about better judgment at work needs a test of judgment at work. Choose a transfer check that resembles the intended use, including the tools and time pressure involved.

For your own learning, spend part of a practice session choosing which method fits. Interleaving different problem types creates opportunities for that choice. Use retrieval practice to bring the method to mind without the lesson in view. Then try a fresh case without a topic label or a hint that gives away the connection.

06Where it doesn’t carry over

Earlier learning can also interfere. This is called negative transfer: an old habit or rule makes the new task harder. Familiar software shortcuts may trigger the wrong command in a different program. A negotiation tactic that works with a one-time seller may damage a continuing relationship. The old response needs a check against the new setting.

A failed transfer test also has several possible explanations. The original learning may be weak, the relevant knowledge may be hard to retrieve, or the new task may require an additional skill. Check whether the person can still perform the original task before diagnosing a transfer problem. Research supports transfer under identifiable conditions; sweeping claims that one exercise improves every kind of thinking need much stronger evidence.

07Roots

Edward Thorndike and Robert Woodworth approached a question with large consequences for schooling: did exercising one mental ability strengthen the mind more broadly? Around 1901, supporters of mental discipline defended subjects such as Latin partly on the grounds that demanding study would improve general reasoning.

The psychologists tested how improvement spread between tasks. Their experiments included estimating the area of rectangles and checking performance on other shapes and sizes. The gains depended on what the training and test tasks shared. Thorndike developed this into an account based on identical elements: learning carries over through components common to both situations.

Later researchers widened the account. Common elements could include a relationship or strategy, and people still had to notice the connection. Gick and Holyoak’s fortress and tumor problems made that recognition problem vivid. The idea traveled into education and workplace training, where researchers now ask both what overlaps between tasks and how learners retrieve, recognize, and adapt what they know.

08How solid is this?

Transfer is well established, especially when tasks share relevant features or structure. Its extent depends on mastery, retrieval cues, recognition, and differences between settings; broad improvements across unrelated abilities are much less reliable.

09Connections

Often confused withLearning Curve
Can follow from Feynman Technique, Interleaving
IncludesNear vs. Far Transfer
See also Deliberate Practice, Working Memory, Testing Effect, Analogical Reasoning, Encoding Specificity, Worked-Example Effect, Peter Principle

+ 1 more in the list

10Origin and sources

Edward Thorndike and Robert Woodworth’s 1901 experiments helped establish the modern study of transfer by testing how improvement in one mental function affected performance in others.

Suggest an edit· Updated 2026-10-02