The Transfer of Learning

We hardly need experimental data to verify that the acquisition of new behaviors is influenced by previous learning. You could not be learning from this text unless you had first learned to read. Yet, your ability to learn this text is not at all affected by your having previously learned to ride a bicycle. What then, are the factors that control the extent to which previous learning will affect new learning?
As you read through the text, consider the following questions.

The effect of previous learning on new learning, either in a facilitatory or inhibitory manner, is called transfer of learning. Formal education is premised on the assumption that what we learn in the classroom will somehow be transferred to the "real" world. In the design of school curricula, course sequences are designed on the assumption that the more elementary courses will aid learning in the more advanced courses. The setting of prerequisites for mathematics or physics is also based on this notion. Experiments have shown, however, that sometimes the expected transfer does not occur.

There is little transfer between dissimilar learning experiences

It was once assumed that taking courses such as Latin or geometry would lead a person to think more logically, and hence benefit the study of science. This assumption has been called the "theory of formal discipline." Thorndike (1924) studied the problem of formal discipline and summarized his conclusions as follows:

The expectation of any large difference in general improvement of the mind from one study rather than another seems doomed to disappointment. The chief reason why good thinkers seem superficially to have been made such by having taken certain school studies is that good thinkers have taken such studies, becoming better by the inherent tendency of the good to gain more than the poor from any study. When the good thinkers studied Greek and Latin, these studies seemed to make good thinking.... If the abler pupils should all study Physical Education and Dramatic Art, these subjects would seem to make good thinkers.

Thorndike's work had a great impact on educational practices. The theory of formal discipline was overthrown. Thorndike continued his study of transfer, and eventually formulated the "theory of identical elements." This theory states that previous learning facilitates new learning only to the extent that the new learning task contains elements identical to those in the previous task. though the theory of identical elements does not account for all transfer effects, the data supporting its importance are convincing.

The best way to understand transfer is to examine the way we test for it. The most frequently used method requires some subjects to learn task A, then task B, while a control group learns only task B. We measure the difference between the groups' ease in learning task B.

Experimental Group Learn Task A Learn Task B
Control Group Learn Task B

When we analyze the learning of task B, there are three logical outcomes.

a. The experimental group could learn it more easily than the control group. This would indicate positive transfer.

b. The control group might learn task B more easily than the experimental group. This would indicate that task A somehow inhibited the learning of task B. This would be called negative transfer.

c. The experimental and the control groups could do equally well in learning task B; no transfer.


Several measures of performance have been used in transfer studies.
These are:

a. the number of trials required to reach some level of mastery,
b. the level of learning attained after a fixed number of trials, and
c. the number of errors made before mastery is reached.

Once the data have been collected, they are usually summarized by a
transfer equation. When the measure used is the number of trials to
mastery, the percentage of transfer = (C--E) / C times 100, where E is the mean
performance of the experimental group and C is the mean performance of
the control group. Multiplying by 100 only serves to convert the ratio to a percentage. You can leave that step out and still get the picture

If the following data were obtained:

Mean Trials to Mastery
Experimental Group6
Control Group10

Then the percentage of transfer (10-6)/10 X 100 = 40%

We could say that there was a transfer of 40% from the first task to the second. If, however, the data appeared as follows:
Mean Trials to Mastery
Experimental Group15
Control Group 10

Then the percentage of transfer = - 50%

There is a negative transfer of 50% from the first task to the second. Learning the first task hinders the learning of the second task.


As previously indicated, if there are identical or even similar elements in two tasks, transfer may occur. There are three ways in which tasks can be similar. These are:
a. stimulus similarity,
b. response similarity, and
c. procedural or principle similarity.

Each of these components can be broken down into two types: similarity in form or similarity in meaning.

The formal properties of a stimulus are its physical dimensions. They are size, shape, color, sound, taste, number, and so on. Fat and cat are similar in form because they both are three-letter words, and both contain the A and T in the same relative positions.

Two items are similar in meaning if they have similar definitions or associations; thus, cathedral and church would be considered as having high similarity. Cathedral and bell would be judged as dissimilar. Many methods have been used to measure similarity in meaning; most involve some form of associative strength.


If the stimulus and response portions of tasks A and B are identical, we would expect a high degree of transfer. But what happens if one or the other, or both, are different?

Poffenburger (1915) attempted to answer this question. He did a series of four experiments designed to compare the effects of stimulus and response variables on transfer. The experimental designs and results are summarized below.
GroupTask ATask BSimilarity/Differences in Task B
1S1-R1S1-R1 Same stimulus.. Same response. High positive transfer.
2S1-R1S1-R2 Same stimulus Different response. Negative transfer.
3S1-R1S2-R2 Different stimulus Different response. Little or no transfer.
4S1-R1S2-R1 DIfferent stimulus. Same response. Posi- transfer.

Note that in every group all subjects learned the same task A, but they learned a different task B, depending on which group the subjects were in. Let us analyze each of these groups.

Group 1 Because this group was given the same stimuli and responses in both tasks, there was high positive transfer.

Group 2 -- In this group the same stimuli were used, but different responses had to be associated with them. An example of this occurs when introducing an old friend who has just married; she is no longer called Mary Boone, but is now Mary Smith. Here, all the stimuli are the same -- her appearance and her first name -- but the final response, last name, is different. The response "Boone" must be unlearned during the acquisition of the new correct response, "Smith." Negative transfer is the result.

Group 3 -- This group was given a new task with different stimuli and with different responses. Since there is no similarity between either the stimuli or responses of the tasks, we might expect no transfer. But when there are similarities in the procedures used in each task, we often find some slight positive transfer anyway.

Group 4 -- This group has to attach an old response to a new stimulus. In this case, the two tasks have more similarity than they did for Group 3, and there is no need to unlearn responses, as Group 2 had to do. The result is positive transfer.

Bruce (1933) first had his subjects learn of nonsense paired-associates and then learn a second list which varied either in the stimulus, the response, or both. Sample items are shown below:
Sample Task A Sample Task B% Transfer
Group 1 XAL -- BOM CAM -- LUP (both different) 16%
Group 2 LUN -- GIP FIS -- GIP (stimulus different, response same)37%
Group 3 REO -- KIV REO -- ZAM(stimulus same, response different)-9%

The 16% transfer shown by Group 1 is due to similarity in procedures used, (sometimes called nonspecific factors). Experimental Group 2 showed that similar responses produce additional positive transfer. In contrast to the facilitatory effects noted in the first two conditions, Group 3 shows that stimuli with different responses produce an inhibitory effect and negative transfer results. This latter fact is an exception to the general rule that similarity facilitates transfer. If the responses are different, then the greater the similarity of stimuli, the less the positive transfer. The inhibitory effect is attributed to the fact that the subject has to unlearn an old response before can making a new response to a stimulus. The major problem in education is to design curricula with maximum transfer to real life


In general we can conclude that to maximize transfer in educational situations we should maximize the similarity between the learning task and the testing or performance situation. To teach one way and test in another way will reduce transfer. Yet, this principle is often overlooked. In industrial training the conflict between book learning and on-the-job training is common. Training experts are aware of this problem and attempt to make the learning situation as "real" as possible. The military spends large amounts of money to build simulators which present the student with as close an approximation of actual situations as possible. The astronauts practiced long hours in simulation equipment prior to their moon landings.

Just how much of one's school career will transfer to later life? Trow (1957) gives his view that the schools do not teach for transfer to real life.

The study of mathematics and science is necessary to produce mathematicians and scientists, true enough. This is the direct, content approach. But will the study of algebra reduce the annual slaughter of 40,000 persons on our highways or will transfer be more likely from courses in driver education? And will five year's study of Physical Sciences serve to reduce the number of resident patients, now about 500,000, in our mental hospitals, or the 2,500,000 major crimes committed in this country each year?

Using the principles of transfer of training we have Just studied, what sorts of courses would you recommend to help solve these social problems?


Now test yourself without looking back.

1. Previous learning can act to inhibit the learning of later tasks. This phenomenon is called________________________________________.

2. Which condition would produce maximum facilitation?
Task A Task B
a. S1 - R1S2-R1
c. S1-R1S2-R2

3. In which condition would we attribute any positive transfer resulting to nonspecific factors?
Task A Task B
a. S1 - R1S2-R1
c. S1-R1S2-R2

4. By formal similarity we mean:
a. the stimuli look similar.
b. the stimuli have common associations.
c. (neither)

5. An industrial psychologist wants to design a teaching procedure that will hopefully transfer to the job situation. What general rule should be followed?____________________________________________________________

6. The theory that studying Latin transfers to solving problems in analytic geometry is called the theory of______________________________________________

7. The following are the results of a transfer experiment. Compute the percent of transfer.
Mean Trials to Mastery
Experimental Group8
Control Group12



Below is a typical transfer design.
Experimental Group Learn Task A Learn Task B
Control GroupLearn Task B

The effects of having learned Task A on the learning of Task B may be facilitatory or inhibitory or none at all. Match the effect with the possible outcome of the experiment.

1) Experimental group makes fewer errors in the learning of Task B than does the Control Group._______
2) Experimental Group makes more errors in the learning of Task B than does the Control Group._______
3) Experimental Group requires fewer trials in the learning of Task B than does the Control Group.________
4) There is no difference between the errors made by either group._________

a Facilitates (Positive transfer)
b. Inhibits (Negative transfer)
c. No effect (No transfer)

If the Control Group's performance is superior to the Experimental Group's, negative transfer is said to have occurred. This means:

a. Task A facilitates the learning of Task B.
b. Task A inhibits the learning of Task B.
c. Task B inhibits the learning of Task A.
d. (none of these)

_____________________________________________ 4

The amount of transfer produced can be computed by means of a transfer formula. One such formula is: % transfer = (C- E)/C times 100

Use this formula to compute the percentage of transfer for the following experiments.

a. Experimental Group
Control Group
Mean Trials To Mastery Task B

b. Experimental Group
Control Group
Mean Trials to Mastery Task B


There can be transfer between two tasks only if there is some similarity between the two tasks. Transfer may occur if:

a. the stimuli are similar.
b. responses are similar.
c. the procedures or principles are similar.
d. (none of these)

__________________________________________ 5

One way two stimuli can be judged similar is called formal similarity. Formal properties of stimuli are those that can be directly sensed: shape, tone, quantity, and so on. Which of these are formal properties?
a. Size
b. Meaning
c. Understanding
d. (none of these)

_________________________________________ 3


a. --67%
b. 50%

1) a
2) b
3) a
4) c

3. a
4. b
5. a,b,c

A second way stimuli can be similar is if they either have associations in common or are often associated with one another. Thus, husband is judged similar to wife and house similar to home. The husband-wife similarity is based on the fact that:

a. people tend to marry from similar backgrounds.
b. one is often given as an association of the other.

If we vary the similarity of stimuli and responses independently, we find the following results:
Task ATask B
S1 - R1S1-R1Same stimuli, same response -- high positive transfer
S1-R1S1-R2 Same stimuli, different response -- negative transfer
S1-R1S2-R2Different stimuli, different response -- little or no transfer
S1-R1S2-R1 Different stimulus, same response -- positive transfer

Predict the transfer (either +, or -, or little or none) in the task below.
Task ATask B
a.LAX -- MIP MAC -- PUL________
b. NUL -- PIV GIF -- PIV______
c. GER -- VIKGER -- MAZ _________


Frequently, positive transfer is noted even though there is no similarity between either the stimuli or responses. This is attributed to similarity in procedures. In which of the experiments below might positive transfer be produced?
a. Task ATask B
b. Task ATask B

__________________________________________ 4

The theory of formal discipline predicted that such courses as Latin and geometry would build an analytical mind and thus facilitate later learning of such courses as physics. When the theory of formal discipline was tested, it:
a. was not supported by experimental evidence.
b. was upheld in the laboratory.

_________________________________ 1

Which reason most likely explains why the astronauts practiced in exact replicas of the command module prior to a trip to the moon?
a. Because the closer the training situation is to the test
b. Because general principles of operating modules could apply to a wide variety of such space craft.


1 a
2 b
3 a
4 b -
5 a.little
b. + transfer
c. - transfer


1. Which group(s) would show negative transfer in the experiment summarized below?
Task A Task B
a. S1 - R1S2-R2
c. S1-R1S2-R2

2. The theory of formal discipline would predict that:
a. stimulus similarity is important in transfer.
b. taking geometry transfers to later tasks that include analytical thinking.
c. similarity in form is the most important factor in transfer situations.

3. Similarity in meaning indicates that:
a. the principles in the two tasks are identical.
b. the stimuli s sounds are similar.
c. (neither)

4. The fact that some positive transfer often occurs from Task A to Task B even though neither the stimuli or responses are similar is attributed to____________________________________

5. What type of transfer would result from the following condition?
Task aTask B
S1 -- R1S2 -- R1

6. The following are the results of a transfer experiment:

Mean Trials to Mastery
Experimental Group8
Control Group5

Compute the percent transferred.


Click here to return to Table of Contents Module 4