How People Learn‎ > ‎

03. Learning and Transfer

        Transfer, the ability to apply learning from one context to another, is a basic educational assumption that can take different forms.  Thorndike first applied transfer to building a "mental muscle" that can be applied to many situations; he looked for common elements within tasks.  Today this model is intact, however it has been extended to consider types of practices and learner characteristics.  Key characteristics of transfer include the importance of having initial learning to transfer, certain types of learning experiences that support transfer, abstractions, being best viewed as an active and dynamic process, and understanding that new learning involves transfer from previous learning.     
        For transfer to take place, there needs to be a threshold of initial learning, dependent on the degree of understanding.  Reaching this depends in part on time spent, as without it students can have difficultly reaching understanding.  The time actually spent is also important to this.  Learning is effective when people engage in "deliberate practice" - monitoring one's learning as they learn.  Understanding when, why, and where to use knowledge is effective too, as with "contrasting cases".  Further, transfer is enhanced when students see the potential for transfer.  
        Motivation, both intrinsic and extrinsic, is important to how much people are willing to learn.  Tasks must be at the appropriate level of difficultly - not easy and boring, nor too hard and frustrating.  Learners' can have tendencies too, specific to a subject.  "Performance oriented" students are more worried about making errors, while "learning oriented" students like new challenges.  Social factors, such as feeling that one is contributing to others, and seeing the usefulness of the task are motivating as well. 
        Context is an important factor for transfer too, as for example street vendors can perform mental arithmetic well on the street yet struggle in formal settings.  When material is taught in multiple contexts, it is more likely to be transferable.   When material is overly contextualized, wide transfer can be problematic.  Methods to overcome this include solving similar cases to abstract general principles, to allow students to engage in "what-if" questions, and finally to generalize that case, so that instead of finding a solution to one problem they should find it to a set of problems.  Representing problems abstractly has been shown to enhance transfer.  
        Identifying elements of transfer is important in determining what is actually being learned.  There is support that cognitive elements is the common link between tasks.  There is also support that teaching abstract representations become components of schemata.  
        Transfer should be viewed as dynamic, not static.  Instead of measuring if calculus helps on day 1 of physics, for example, it should be viewed on quickness to learn throughout the year.  Sometimes, prompting is important for transfer to occur.  Graduated prompting is a dynamic assessment method that  grew out of this view.  
        Metacognitive tasks where students actively monitor their strategies have shown positive transfer results.  One is reciprocal teaching, where students take turns reading and teaching comprehension strategies to one another.  Procedural facilitation and Schoenfeld's heuristic methods for mathematical problem solving have similar aspects, involving making students ask metacognitive, self-regulatory questions.   New technologies that allow modeling open opportunities for performing such metacognitive tasks.  
        "All learning involves transfer from previous experiences."  There are 3 major implications of this idea.  First, it is important to build on existing knowledge.  Oftentimes students have well-developed understandings but if the new material fails to build on it, it can be lost.  Second, previous understanding may be wrong, and without addressing it can lead to further wrong assumptions.  Third, teachers much try to make students' understanding visible and help reconceptualize faulty ideas. 
        Prior knowledge and behavior is context-sensitive.  For example, while Anglo mother and African American mothers have the same number of language interactions with their children, the latter focus more on affective dimensions.    It is important for educators to build on, rather than view certain cultural habits as deficits.  
        Transfer into everyday life is important to consider, particular if education is aimed to be practical.  However, contrasts exist between school and everyday life.  These include more collaborative work in everyday life, the use of tools in the real world as opposed to mental tasks in school, and more contextualized reasoning in everyday settings.  These are points that ought to be considered, not necessarily used, as they contradict other research (such as teaching in overly contextualized environments not leading to transfer).  What is suggested is that a balance of specific and general principles should be balanced with learning tasks.  
        In summary, transfer is important because education tries to prepare students for flexible adaptation to novel problems.  Teaching needs to consider prior knowledge, students must monitor their own learning, context must be considered, and there should be a balance of general and specific principles.  All new learning involves transfer.  Transfer to everyday environments is the ultimate purpose of school. 

Initial Response to the topic: 3 thoughts/ideas, 2 questions, 1 analogy

        Transfer is the purpose of all education.  I think it is the most fundamental aspect of any learning setting.  To really test learning, students should be tested on novel problems to see if they can transfer over the ideas.  
        What would be the point of learning if we cannot apply the learned concepts to another setting?  It seems very hard to teach transfer - how do we know the new problem has similar concepts as the original ones?
        Transfer is to education as gas is to a car.  

New Responses to the topic: 3 thoughts/ideas, 2 questions, 1 analogy

        The fact that everything we do or know involves transfer from previous knowledge shows why it is so fundamental to learning.  It makes the implications for the need for students to monitor their own learning clear, as they must actively connect what they understand to the new situation.  Moreover, that concepts taught should be both contextualized and abstracted is an interesting yet important way to analyze teaching material.  Pulling out the abstract principles from a contextualized problem seems to be an important learning task.  Finally, that education's purpose is to make lifelong learners and independent thinkers, which is essentially that learners can approach new situations with and flexibly adapt, is the best articulated educational purpose known to me.  
        The questions that I have are practical.  How do we design curriculum to consider what we know about transfer?  This seems to change the "templates" that we use, but in what way?  Also, how do we assess transfer?  Is it just as simple as assigning novel problems to students that have similar general principles?  I am certain it is complex and needs relatively deep thought.  
        An analogy that I would draw is one to Bloom's taxonomy, that has application, analysis and synthesis at the top levels.  Transfer seems to fit in very well with application, though perhaps it suggests that analysis and synthesis are below it on the order of learning, or perhaps can just be categorized in a non-hierarchical framework.