May 2014: Using Cognitive and Educational Psychology to Improve Learning

Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4–58.

Summary by Dr. Paul Cirino


This monograph by Dunlosky, Rawson, Marsh, Nathan, and Willingham (2013) reviews research-based principles and techniques from cognitive and educational psychology that have been linked to improved student learning. The authors evaluate 10 different learning techniques in a detailed, structured manner that includes the mechanism by which each operates, the situations in which the technique is effective, and potential issues surrounding implementation. The authors focus on techniques that are known to be effective, and/or known to be used frequently, and do not require extensive training. The authors provide a definition of each technique in Table 1 (p. 6) and the conditions that could affect generalizability or otherwise influence effectiveness in Table 2 (also p. 6). 

The authors conclude that the 10 techniques have varying utility (which includes how well they work and their generalizability): high, moderate, and low. The two techniques rated “high” are distributed practice and practice testing. The three techniques rated “moderate” are interleaved practice, elaborative interrogation, and self-explanation. The remaining five techniques (summarization, rereading, highlighting and underlining, keyword mnemonic, and imagery for text) are given “low” ratings. In Table 4 (p. 45), the authors provide a summary of qualitative effectiveness ratings. All three tables are reproduced below.

Table 1 Table 2 Table 4

1. Elaborative interrogationGenerating an explanation for why an explicitly stated fact or concept is true
2. Self-explanationExplaining how new information is related to known information, or explaining steps taken during problem solving
3. SummarizationWriting summaries (of various lengths) of to-be-learned texts
4. Highlighting/underliningMarking potentially important portions of to-be-learned materials while reading
5. Keyword mnemonicUsing keywords and mental imagery to associate verbal materials
6. Imagery for textAttempting to form mental images of text materials while reading or listening
7. RereadingRestudying text material again after an initial reading
8. Practice testingSelf-testing or taking practice tests over to-be-learned material
9. Distributed practiceImplementing a schedule of practice that spreads out study activities over time
10. Interleaved practiceImplementing a schedule of practice that mixes different kinds of problems, or a schedule of study that mixes different kinds of material, within a single study session
MaterialsLearning conditionsStudent characteristics aCriterion tasks
VocabularyAmount of practice (dosage)AgeCued recall
Translation equivalentsOpen- vs. closed-book practicePrior domain knowledgeFree recall
Lecture contentReading vs. listeningWorking memory capacityRecognition
Science definitionsIncidental vs. intentional learningVerbal abilityProblem solving
Narrative textsDirect instructionInterestsArgument development
Expository textsDiscovery learningFluid intelligenceEssay writing
Mathematical conceptsRereading lags bMotivationCreation of portfolios
MapsKind of practice tests cPrior achievementAchievement tests
DiagramsGroup vs. individual learningSelf-efficacyClassroom quizzes

a Some of these characteristics are more state based (e.g., motivation) and some are more trait based (e.g., fluid intelligence); this distinction is relevant to the malleability of each characteristic, but a discussion of this dimension is beyond the scope of this article.

b Learning condition is specific to rereading.

c Learning condition is specific to practice testing.

TechniqueUtilityLearnersMaterialsCriterion tasksIssues for implementationEducational contexts
Elaborative interrogationModerateP-IPIPI
Keyword mnemonicLowQQQ-IQQ-I
Imagery use for text learningLowQQQ-IPI
Practice testingHighP-IPPPP
Distributed practiceHighP-IPP-IPP-I
Interleaved practiceModerateIQP-IPP-I

Note. A positive (P) rating indicates that available evidence demonstrates efficacy of a learning technique with respect to a given variable or issue. A negative (N) rating indicates that a technique is largely ineffective for a given variable. A qualified (Q) rating indicates that the technique yielded positive effects under some conditions (or in some groups) but not others. An insufficient (I) rating indicates that there is insufficient evidence to support a definitive assessment for one or more factors for a given variable or issue.

The authors also identify areas that remain unclear, such as (a) how individual differences (e.g., age or level of ability) might influence effectiveness of a given technique; (b) how long a given technique exerts its effects; and (c) how transportable the results might be to the other contexts, including classrooms and coursework. The authors suggest future work in these areas.

Key Findings

Distributed practice (high). Distributed practice refers to distributing or “spacing” studying and learning over time, rather than cramming or “massing” the same information in back-to-back sessions or within a short time period. The technique also refers to “lag effects,” where spacing with longer intervals is better than spacing with short intervals. This type of relearning is thought to affect learning via a number of mechanisms, including the fact that greater effort is required for retrieving previously learned information when the next opportunity for learning happens farther away in time from the original learning (but not too far away in time). Distributed practice is effective across individuals of different ages and skill levels and for different types of materials (e.g., vocabulary, science, math) and criterion tasks (e.g., free recall, problem solving, achievement tests, classroom quizzes). It is suggested that the intervals between original learning and practice or relearning should be approximately 15% of the desired retention interval (e.g., to retain something for 150 days, separate learning sessions by 10 days). Interestingly, some evidence suggests that students are not necessarily aware of the benefits of distributed practice, even when they have benefited from it, and so training may need to include a progress-monitoring component before students use distributed practice techniques spontaneously.

Practice testing (high). Practice testing includes instructor- or student-initiated testing. Examples include practice recall of information by using flashcards, completing end-of-chapter practice quizzes or problems, or completing exercises included in supplementary materials. This technique is explicitly distinguished from the high-stakes standardized testing that often occurs in schools and for other educational purposes. One way that this technique improves learning is by helping students to mentally organize to-be-learned material. Effects are consistent and generalizable across the learning conditions that the authors examined but are most robust when (a) responses are constructed, such as short answers or longer responses (as opposed to, for example, recognition); (b) dosage is increased (more is better, with and across practice sessions); and (c) timing is distributed (spaced practice is better than massed practice, within and across practice sessions). All age ranges have been shown to benefit, though skill level, such as whether the learner is typically developing or has more difficulty in learning, has been much less studied. A weakness in the literature is that the most common learning materials involve stimuli, such as paired associate word learning, individual facts, and definitions, though effects have been shown for more complex materials, such as texts of varying length and video lectures. Practice testing is most often assessed through recall of the same information that was learned earlier, though effects are noted even when only a sample of the earlier learned material is tested or when criterion tests involve inference. Effects are most common over very short retention intervals but are also noted over retention intervals ranging from weeks to months. Practice testing with corrective feedback yields the highest rates of learning.

Interleaved practice (moderate). Interleaved practice is a schedule of practice that combines different types of problems or materials within a single study session, as opposed to blocked practice in which all content for a particular unit, topic, or problem type is studied before moving to something different. A typical result is that studying in blocks is better than interleaving during initial learning and practice, but studying by interleaving is better than blocking at a later time point. Interleaving is suggested to work by (a) allowing for discrimination between problem types more readily, which helps memory, and/or (b) forcing retrieval from long-term memory more often than  blocking. Interleaved practice is related to though separable from distributed practice, as material can be spaced without being interleaved with different content, and some research has demonstrated the effectiveness of interleaving when spacing is controlled. Interleaved practice is also relatively easy to implement. A downside is that the literature in this area is more robust for motor than cognitive learning, and it is as yet unclear how much blocked practice is required (to attain competency) before moving to interleaved practice (though more is suggested for younger or less-skilled learners). Evidence appears mixed for a variety of students and types of materials (and seems most consistent for math) and for different retention intervals.

Elaborative interrogation (moderate). Elaborative interrogation involves stating a fact and asking “why” questions in a compare-and-contrast format. Elaborative interrogation is thought to assist in the connection between prior knowledge and current information in text or a lecture. Elaborative interrogation has been studied primarily outside of the classroom and for students in upper elementary and above, with most research in undergraduate students. The assessment of whether elaborative interrogation helps learning is done via recall or memory tasks and mainly over short intervals (though some research shows that retention of learning over several days seems promising when using this technique). Elaborative interrogation seems most effective when prior domain knowledge is high rather than low, presumably because there is more information with which to tie newly learned information. Elaborative interrogation is effective for discrete fact knowledge. However, there is a strong need for evidence of effectiveness with a wider range of to-be-learned material.

Self-explanation (moderate). Self-explanation involves describing how new information in text or a lecture relates to information the student previously studied or already knows. The explanation of why this technique works has to do with the processes that occur while learning new material. Similar to elaborative integration, self-explanation helps the student to connect prior and current information, which leads to new learning and better retention of old and new information. The prompts used to trigger self-explanation range from content-specific to more abstract. The current review focuses on the latter and includes general prompts such as “What new information is provided?” and “How does this relate to what you already know?” This technique has been demonstrated to work across grade levels and with a wide variety of learning materials and criterion measures. However, the role of domain knowledge is not clear. Further, most studies examine only the immediate effects of single-session learning, and examination of its utility in classroom settings is very limited. Some training may be required to effectively model and teach students how to engage in effective self-explanation, rather than just paraphrasing.   

Summarization (low). Summarization typically involves summing the contents of a passage. It is thought to work by helping the learner identify and organize the higher-level meaning of a passage. One of the weaknesses in this literature is that what counts as a summary is highly variable (written vs. spoken, with vs. without the passage, of differing lengths), such that general conclusions are elusive. The quality of the summary seems paramount—learning is maximally benefited when the information of the summary is correct and connects to prior knowledge. Benefits have been shown for older students but can also be observed for middle school students with sufficient training. The feasibility and effects on learning of summarization in younger students is much less clear, and more generally, individual differences in summarization skill  affect its effectiveness as a study technique. Studies of summarization occur in classroom settings more often than in the laboratory, which speaks to their feasibility. Text passages are most frequently studied, though the impact of text characteristics (e.g., readability, length, structure) is not well specified. When effects are found, they can be durable on the order of days to weeks. The authors differentiate between summarization as a study technique (the focus of this monograph) and using summaries to check understanding (e.g., as an outcome themselves). A key issue regarding the effectiveness of summarization is that it appears to be consistently less strong relative to generating explanations or to self-questioning. In other words, when instructing and encouraging students to use various learning techniques, generating explanations and self-questioning should be given precedence over summarization.

Rereading (low). Rereading text is very popular, direct, and easy to do at a surface level. It is presumed to work both because the reader encodes information more than once (the quantitative hypothesis) and because rereading is thought to improve conceptual organization and extraction of main ideas (i.e., more resources can be devoted to these higher-level processes during rereading than during initial reading—the qualitative hypothesis). In general, the latter hypothesis is more often supported (rereading boosts recall of main ideas more than recall of the details of a passage). Spaced rereading tends to work better than massed rereading, and effects are shown for a variety of learning materials. However, little is known regarding how individual differences (e.g., prior knowledge, general skill) affect rereading, and nearly all studies focus on undergraduates. There is mixed evidence regarding the persistence of rereading effects over time. Interestingly, effects are more often found for free and cued recall, rather than multiple-choice questions. A main issue with rereading is not whether it is ineffective, but that it appears to be less effective when compared directly to other learning techniques (e.g., elaborative interrogation, self-explanation, practice testing).

Highlighting and underlining (low). Like rereading, highlighting is frequently used, although evidence for its specific benefit (i.e., when not paired with other techniques on this list) is lacking. The idea is that information that is highlighted or underlined will more easily capture the reader’s attention upon review and presumably lead to greater recall. The key is that the effective components depend on the students’ ability to determine which part of text is the most important. Because highlighting and underlining draw attention to specific text, it may work better for outcomes that are fact based than for those that require inference across the passage. The authors suggest that it may be worthwhile to train students how to efficiently highlight or underline the most important information, if only because students frequently use highlighting and underlining and will likely continue to do so, despite its lack of utility in general.

Keyword mnemonic (low). Keyword mnemonic refers to a form of mental imagery, which has engendered a large body of literature in its own right. Here, students associate a to-be-learned word or proposition with a more accessible one, and then they create an interactive image pairing the two (an example is given of a dentist holding a large molar with pliers to remember that the French la dent means “tooth”). Keyword mnemonic is useful for easily imaged words (concrete nouns) but much more difficult to use for abstract terms, and upon retrieval, the key referent may be lost (the example given is whether la dent means “tooth,” as opposed to “molar,” “pliers,” or “enamel”). Also, where useful, studies typically provide the associational material, which requires time and effort to produce. Finally, any benefits that show retention over time have been in cases where testing occurs after an immediate delay (this process is closely associated with practice testing; see above).   

Imagery for text (low). Like keyword mnemonic, this technique involves imagery; the difference is that here, the imagery refers to the content of a text, rather than word associations. Imagery for text is relatively easy to implement. However, imagery’s effectiveness is in general mixed and appears to depend on a range of factors, including (a) the type of text (whether its content can be easily depicted as images), (b) how the text is delivered (more effective for text that is listened to than for text that is read), and (c) the extent to which students actually use imagery (students may use it to varying degrees, whether or not they are in an “imagery” condition or a “nonimagery” condition). Because the technique takes place in the student’s head, it is difficult to know how much imaging the student is using and whether the imagery is relevant to what the student is reading or listening to. It is unknown whether effects found for this technique are retained over time, as few studies have investigated this technique, and most studies have not shown transfer to authentic educational text or standardized tests.

Conclusions and Recommendations

  1. This monograph communicates to teachers and students which learning techniques are evidence based. It provides needed information about high-yield techniques for optimal learning and information about techniques that are widely used but have less evidence of effectiveness (e.g., highlighting).
  2. Students may not spontaneously use a beneficial technique (e.g., distributed practice). Students may need to be directly instructed in the technique, reminded of its effectiveness, and monitored with regard to its implementation over time.
  3. A key feature of most of the effective techniques is that they require active processing and/or retrieval of information. For some techniques (distributed practice and interleaved practice), timing appears to be as important as the content or learning activity itself. Incorporating the most effective techniques reviewed here (practice testing and distributed practice) into current classroom curricula is likely to benefit students, particularly if done properly—that is, efficiently, with appropriate spacing, and with appropriate training and/or feedback. Notably, many textbooks group problems of the same content or unit and do not come back to old content or problems in later units. Therefore, educators may need to build in their own instructional prompts to explicitly schedule interleaved or distributed practice of newly learned material over time.
  4. The effectiveness of these learning techniques has primarily been demonstrated with older students (e.g., undergraduates), so there is some uncertainty regarding the relative effectiveness for elementary-grade students. This fact is relevant for at least three reasons. First, there is relatively little information regarding how the techniques work at different levels of skill and prior knowledge, of which younger students have less. Second, younger students may find some techniques (e.g., summarization) difficult to implement without extensive training and practice. For these students, the metacognitive and motivational requirements of the techniques, alone or in combination, are also not well delineated, and therefore, studies that approach such questions are likely to enhance our knowledge in this area. For example, comparing or combining activities that operate similarly (e.g., elaborative interrogation and self-explanation both enhance connections between current and prior knowledge) may help distinguish which is more effective and how the effectiveness varies with student characteristics such as grade and level of background knowledge. Third, few reviewed studies focused on individuals with learning disabilities. Given the pressing need for both early intervention and continued work with students with learning disabilities throughout their schooling, the extent to which these techniques might operate in these populations is important to know. If we assume that learning disabilities represent the lowest end of a continuous distribution of skill (as research is beginning to show), then it is reasonable to hypothesize that techniques that are effective for typically developing learners might also work for struggling learners. However, such techniques may require additional instruction and practice for students with learning disabilities.