The Center for Teaching and Learning at UC Berkeley breaks down some foundational principles and research about memory and recall; the Science of Learning; and evidence-based teaching strategies, such as elaborative interrogation, self-explanation, practice testing, distributed practice and interleaved practice.

Before we dive into understanding the relevant science behind the learning process, let’s ground ourselves in a definition of learning that is drawn from research.

Learning is a process that:

• is active: process of engaging and manipulating objects, experiences and conversations in order to build mental models of the world (Dewey, 1938; Piaget, 1964; Vygotsky, 1986). Learners build knowledge as they explore the world around them, observe and interact with phenomena, converse and engage with others, and make connections between new ideas and prior understandings.
• builds on prior knowledge and involves enriching, building on and changing existing understanding, where “one’s knowledge base is a scaffold that supports the construction of all future learning” (Alexander, 1996, p. 89).
• occurs in a complex social environment and thus should not be limited to being examined or perceived as something that happens on an individual level. Instead, it is necessary to think of learning as a social activity involving people, the things they use, the words they speak, the cultural context they’re in and the actions they take (Bransford, et al., 2005; Rogoff, 1998), and that knowledge is built by members in the activity (Scardamalia & Bereiter, 2006).
• is situated in an authentic context: provides learners with the opportunity to engage with specific ideas and concepts on a need-to-know or want-to-know basis (Greeno, 2006; Kolodner, 2006).
• requires learners’ motivation and cognitive engagement to be sustained when learning complex ideas because considerable mental effort and persistence are necessary.

The conditions for inputs to learning are clear, but the process is incomplete without making sense of what outputs constitute that learning has taken place. At the core, learning is a process that results in a change in knowledge or behavior as a result of experience. Understanding what it takes to get that knowledge in and out (or promote behavioral change of a specific kind) can help optimize learning.

Foundational Principles

Humans cannot process everything they experience at once, and if it is not processed, it cannot be learned.

What percentage of time during a lecture do you think students are paying attention?

The fact is you don’t have (much of) their attention. Observational and student self-report studies suggest that during lectures, college students are paying attention 40 to 65% of the time.

Ever feel like you are always having to repeat yourself?

One reason is due to the percentage of time students are paying attention. A second reason is that retaining information, even when paying attention, is a really hard thing to do by itself. In fact, testing immediately following a 50-minute lecture suggests retention rates of 40 to 50%. More so, as the length of a lecture increases, attentional lapses become more frequent, and as a result the proportion of material remembered decreases further.

If paying attention is so difficult, and retrieval of information even harder, what’s to be done?

1. Manipulate factors that you can influence to increase attention. There are four documented factors that impact how a learner pays attention: arousal, interest, fluency and enjoyment. Identify ways to generate and facilitate these (mostly) intrinsic motivators, and student attention during class should increase accordingly.

2. Break up the lecture to help students re-focus. Not only does a change of pace allow students to re-start their attention clock—which is on a 15- to 20-minute attention curve—but attention can increase somewhat dramatically by utilizing other pedagogies in addition to lecture in balanced ways. Student reports of attention during discussion are ~75% and during problem solving ~85%. If that wasn’t enough good news to compel a balanced pedagogical approach, there is evidence of increased focus during lecture immediately following such a change-up. Consider how a balanced pedagogical approach that draws on formative assessment could enrich student learning in a course. Tools like a muddiest point question, minute paper, clickers and think-pair-share are just a few examples of how a formative assessment could help break up a lecture, allow students to re-focus and pay better attention, reinforce retrieval and learning, and possibly serve as jumping-off points to drive a discussion or problem-solving activity.

Memory and Recall

Let’s get a handle first on what we know about memory and recall.

There are two basic types of explicit, or declarative, memory. First, short term or working memory. Think of this as the focus of current attention, or what you are actively thinking about right now. Second, long-term memory is broken down further into semantic memory (facts) and episodic memory (specific events).

Within explicit (or declarative) memory, there are three basic stages of memory processing. Encoding is the process of forming new memories. Storage comes next, and is the process of information maintenance. And finally there is the process of gaining access to stored knowledge, referred to as retrieval. For learning to take place, as we categorize it in this sense, it requires that the information that is processed is then committed to memory and that the student can pull it back out when it’s needed (let alone apply it to a novel circumstance through adaptation and abstraction).

By examining each part of memory and recall from encoding to storage to retrieval, we get a clear sense of how best to optimize this process for our students.

To make encoding a powerful process, it’s necessary to recognize that memories are not stored as faithful recordings, like a book you can pull off a shelf to share again as needed in their exact original form. Instead, each new memory is integrated into our existing body of knowledge—coloring and being colored by other memories. Therefore, the most robust memories are formed through elaboration and organization where learners:

1. process the new information as deeply as possible

2. maximize connections with what is already known

3. situate new knowledge into an existing framework.

 

The challenge with storage is that once something has made it into long-term memory, it tends to remain stored but not necessarily always accessible. The challenge here is not one of capacity. In fact, our capacity for storing new memories is essentially unlimited; more so, organized learning appears to create additional capacity. However, the ability to access a given memory typically declines over time, primarily due to interference caused by the acquisition of new, competing memories. Do not let this sour your hope of students remembering what was learned in your class. You can markedly increase the likelihood of them being able to recall a memory at a future time by strengthening it through retrieval.

Retrieval is an active reconstruction process, not a playback of a memory of an event, fact, concept or process. Every time a memory is accessed for retrieval, that process modifies the memory itself—essentially, re-encoding the memory. The good news: Retrieval makes the memory itself more recallable in the future.

How does it work, and work best for learning?

Retrieval is cue- and context-dependent. To reinforce memory through cues, we’re referring to making as many connections as possible with existing memories. The more possible cues available to elicit retrieval, the better.

With regard to context, the more closely matched the retrieval context is with the context in which the memory was encoded, the better. This holds true even if you can only have one context (such as you only have one shot at retrieval; think high-stakes exam as sole measure of learning in a course). But multiple retrievals in multiple contexts are superior for long-term retention (think frequent low-stakes quizzes that are cumulative). Additionally, we can enable students to do this even better by encouraging them to vary the contexts of retrieval (studying method, physical location), which results in more accessible memories.

Optimize Student Learning

The Science of Learning literature is actually quite clear about this. We are best positioned to optimize memory through “desirable difficulties.”

In this sense, learning is about challenge, failure, understanding why something is wrong before getting it right. This is why most of the pedagogical practices we are about to highlight typically meet (at least initially) with student resistance. The traditional methods of highlighting the textbook and using flashcards simply do not have evidence to support their value in promoting learning. Actually, there are neurobiological reasons these do not work well. It “feels” good to get something right, even when “right” means passively turning over the flashcard to read the correct answer. The brain gives itself a hit of dopamine ("Yeah, you got it right!") thereby reinforcing a habit that doesn’t actually optimize learning.

Challenging the memory system strengthens memories. In other words, make your students “own it" through techniques like Elaborative Interrogation and Self-Explanation.

Retrieving information from memory (or even attempting to!) makes that memory more likely to be recalled. Retrieval via self-testing results in better subsequent memory than the equivalent amount of passive studying (such as highlighting, reviewing flashcards or notes). The optimal technique to leverage here is Practice Testing. 

Retrieval practice is much more effective if sessions are temporally spaced rather than massed. In this case, draw on Distributed Practice.

If possible, topics to be learned should be interleaved by type rather than blocked. This tends to be most applicable to topics in which the need to discriminate among ways to approach and/or solve a problem is critical. The technique of choice to reinforce this is Interleaved Practice.

Elaborative Interrogation

Students:

• Learn explicit fact or concept introduced
• Develop an explanation of why it’s true or applies in a particular case. Why does it make sense that…? Why is it true that…? Why would this be true of [Example 1] but not of [Example 2]?
• Show greater performance improvements when they are asked to develop more precise elaborations; have greater prior knowledge; and generate the elaboration rather than using one provided by instructor, text or other source.

Theoretical Rationale:

• Helps students integrate new knowledge with prior knowledge, enabling students to organize knowledge components in relation to each other, thus promoting retrieval.
• Helps students distinguish among knowledge components and identify when they’re relevant.

Generalizability:

• Learning conditions effective in both intentional and incidental uses
• Material used across a wide range of subject-matter areas and disciplines; must be used with facts or established concepts
• Criterion tasks: Relatively few measures used for retrieval; mixed results on studies of effect when measures of comprehension and application are used; little research on effects after gap between learning and testing; most studies in lab

Implementation Issues:

• Clear guidance as to question types to use and teach students
• Less clear on how specific to make questions on complex processes
• With long texts, students must identify facts to target and frequent use may be needed to ensure performance gains

Overall Rating: Moderate Utility (primarily due to the need for further research)

Self-Explanation

Students:

• Solve a problem using if, then statements or other instructions
• Generate an explanation of the processes used to solve the problem

Theoretical Rationale:

• Helps students integrate new knowledge with prior knowledge
• Mechanisms may differ: content-specific or content-free prompts
• Variation in prompts, tasks and measures studied

Generalizability:

• Learning conditions: Effective with direct instruction and discovery learning; effective with concurrent explanation; most effective when no explanations provided to students
• Materials: Effective across subject matter areas and disciplines and across task types
• Criterion tasks: Wide range of measures (improvement of near- and far-transfer); more studies needed on effects with meaningful delay; most studies in labs ( a few in natural settings are promising)

Implementation Issues:

• Effective across subject-matter areas and task types

Overall Rating: Moderate Utility

• Substantial efficacy across disciplines and task types; improves transfer into new contexts
• No conclusive research on effects with delay, on whether training improves efficacy, and on technique vs. time on task.

Practice Testing

Students:

• Engage in low- or no-stakes practice assessments
• Use any of multiple formats (project phases, problems, questions, or tests/quizzes)

Theoretical Rationale:

• Involves generating information oneself rather than encountering it in an external source
• May prompt learners to develop more elaborated links
• May improve mental organization of knowledge components

Generalizability:

• Learning conditions: Practice testing can benefit learners even when practice test and criterion tests use different formats; the greater the use of practice testing, the greater the performance improvement
• Materials: Most studies involve simple facts or concepts and have involved verbal materials
• Criterion tasks: Cued recall used most often (recall for both facts and concepts improves); several studies suggest improvements to comprehension, inference and application; meaningful delays improve retrieval

Implementation Issues:

• Practice testing material often available to students (textbooks, online platforms)
• Repeated testing until correct answers are repeated across study sessions has greatest effect
• Practice testing consistently produces stronger effects than does re-studying
• Instructors can support student's use of practice testing by using low- or no-stakes testing in class

Overall Rating: High Utility

• Extensive research shows the benefits of practice testing across a range of subject materials, test formats, outcome measures, and with meaningful delays between learning and testing, although more research is needed on the effects of learners’ ability levels and prior knowledge.

Distributed Practice

Students:

• Space learning of specific material over time – within a single study session or across sessions; the term “distributed practice” is intended to indicate the benefit of both spaced practice over time and of longer delays between practice sessions

Theoretical Rationale:

• Competing theories: Reduced attention if practice sessions are too close together; reminding prompts retrieval, thus enhancing memory; or later practice sessions benefit from memory consolidation between sessions. Regardless, while greater delay between study sessions increases forgetting between sessions, it increases accurate retrieval at testing.

Generalizability:

• Learning conditions: Most effective when delay between practice sessions totals 10 to 20% of desired recall interval; intentional processing more effective than incidental processing
• Materials: Effective across subject matter areas, disciplines and task types
• Criterion tasks: Most effective with free recall; effects often strongest with greater delay between practice and test; more studies are needed to ascertain any effects on tasks more complex than basic recall; studies in both labs and natural settings

Implementation Issues:

• Potential obstacles: Textbooks not structured to allow distributed practice; research shows that infrequent testing increases students’ tendency to mass practice just before an exam

Overall Rating: High Utility

• Extensive research shows substantial effects with a range of subject matter areas and task types, as well as with long delays; more research is needed on complex material and more cognitively demanding tasks, as well as on the impact of students’ prior knowledge and motivation.

Interleaved Practice

Students:

• Alternate study topics and/or subject matter during study sessions rather than focusing on a single topic or subject matter area

Theoretical Rationale:

• May teach students to distinguish more effectively between problem types
• May improve organization processing and item-specific processing by increasing students’ ability to compare problem types
• May increase instances of retrieval from long-term vs. working memory

Generalizability:

• Learning conditions: Spacing alone does not produce effects as significant as spacing that includes interleaving; more extensive initial massed practice may be required for students with lower ability levels or prior knowledge
• Materials: Studies across subject matter areas and task types show differing results; effective in math, decision-making in complex situations with multiple dimensions (such as medical diagnoses) and conceptualizing artistic styles; ineffective in second-language vocabulary
• Criterion tasks: Always involved mixed-problem sets; effective with delays up to 1 to 2 weeks; studies in both lab and natural settings

Implementation Issues:

• Interleaved practice should be implemented cumulatively as topics are introduced
• While interleaved practice takes more time, the performance improvements offset the cost

Overall Rating: Moderate Utility

• While significant benefits have been shown for math and interleaved practice improves other cognitive abilities, the literature is still small and more research is needed to understand when students have adequate initial understanding and skill.

View additional teaching resources from the Center for Teaching and Learning at UC Berkeley.

Sources:

Adapted from “Learn” resources from UC Berkeley’s Center for Teaching and Learning.

Patricia A. Alexander (1996) The Past, Present, and Future of Knowledge Research: A Reexamination of the Role of Knowledge in Learning and Instruction, Educational Psychologist, 31:2, 89-92, DOI: 10.1080/00461520.1996.10524941

Bransford, John & Vye, Nancy & Stevens, Reed & Kuhl, Pat & Schwartz, Daniel & Bell, Philip & Meltzoff, Andrew & Barron, Brigid & Pea, Roy & Reeves, Byron & Roschelle, Jeremy & Sabelli, Nora. (2005). Learning Theories and Education: Toward a Decade of Synergy.

Dewey, J. (1938). Experience and education. New York, NY: Macmillan.

Dunlosky, J., Rawson, K.A., Marsh, E.J., Nathan, M.J., & Willingham, D.T. (2014). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest.

Faculty Learning Program, Lawrence Hall of Science and UC Berkeley NSF-WIDER Grant

Greeno, J. G. (2006). Learning in Activity. In R. K. Sawyer (Ed.), The Cambridge handbook of: The learning sciences (pp. 79-96). New York, NY, US: Cambridge University Press.

Huddleston, Chelan. Teaching Excellence Colloquium workshop "How Students Learn" (College of Letters and Science, UC Berkeley)

Kolodner, J.L. (2006), “Case-based Reasoning” in R.K. Sawyer (ed.), The Cambridge Handbook of the Learning Sciences, Cambridge University Press, New York, pp. 225-242.

Piaget, J. (1964). Development and learning. In R. E. Ripple & V. E. Rockcastle (Eds.), Piaget rediscovered (pp. 7-20). (Reprinted in Readings on the development of children, by M. Gauvainand & M. Cole, Eds., New York, NY: W.H. Freeman)

The Reinvention Collaborative, Advancing Undergraduate Education in America’s Research Universities (2015 Conference).

Rogoff, B. (1998). Cognition as a collaborative process. In W. Damon (Ed.), Handbook of child psychology: Vol. 2. Cognition, perception, and language (pp. 679-744). Hoboken, NJ, US: John Wiley & Sons Inc.

Scardamalia, Marlene & Bereiter, Carl. (2006). Knowledge building: Theory, pedagogy, and technology. 10.1017/CBO9781139519526.025.

Vygotsky, L. S. (1986). Thought and language. Cambridge, MA: MIT Press.