Definitions, Underpinnings, and Benefits

• Metacognition literally means “cognition about cognition”. Metacognition is defined as awareness and control of thinking and is conceptualized as having distinct facets, which are illustrated in Figure 1 below:
  • Metacognitive knowledge, or awareness (and beliefs) about cognition, which involves:  
    • Declarative knowledge is knowledge about how the mind operates, knowing yourself as a learner, and knowing which cognitive procedures (or study strategies), are best suited for learning.  For example, knowing that learning is boosted by retrieval practice (recalling information from memory) as compared to simply rereading course content.
    • Procedural knowledge is knowledge about how to perform a cognitive-based procedure (or a study strategy). For example, knowing how to best use retrieval practice to improve achievement.
    • Conditional knowledge is knowledge about when and why to use a cognitive-based procedure (or study strategy). For example, knowing that testing during practice boosts performance after long retention intervals but not immediately after practice is conditional knowledge.
  • Metacognitive regulation, or control of cognition, which involves:
    • Planning on how to obtain one’s learning objectives;
    • Monitoring the progress one is making while studying or while taking a test;
    • Evaluating how one prepared for an exam and one’s overall success, such as deciding which of two strategies was more effective at helping a student perform well on an exam.

Metacognitive facets are malleable and can improve either through direct instruction or more organically through experience learning and studying.

• Metacognitive facets are central to effective self-regulated learning, which in general refers to learning that is metacognitively guided by students.  For example, when regulating their learning, students may rely on their metacognitive knowledge to select strategies for achieving a learning goal. They may rely on their metacognitive regulation to monitor how effectively they have acquired knowledge toward achieving their goal, regulate or adjust their strategies, and then subsequently evaluate the quality of their exam preparation.
• Study strategies include the knowledge and beliefs students hold about how to learn course content and prepare for exams.  Accordingly, study strategies comprise an important subset of declarative knowledge.
• A mnemonic tool pertains to activities or strategies that improve students’ learning (or memory) of content, whereas a metacognitive tool pertains to any activities that help students to accurately monitor or control their learning.

Figure 1: Metacognition framework (modified from Schraw & Moshman, 1995)

Flavell, J. H. (1979). Metacognition and cognitive monitoring:  A new area of cognitive-developmental inquiry.  American Psychologist, 34, 906-911. In his classic article, John Flavell introduced foundational concepts of metacognition and proposed “that the monitoring of a wide variety of cognitive enterprises occurs through the actions of and interactions among four classes of phenomena: (a) metacognitive knowledge, (b) metacognitive experiences, (c) goals (or tasks), and (d) actions (or strategies)” (p. 906).  For example, a student may have the goal to complete a worksheet involving word problems relevant to molecular biology, and her metacognitive knowledge may include that she stays more focused in a quiet room (than when music is playing) and that she learns more when using worked examples.  Thus, she decides to use worked examples (a strategy) and monitors her progress toward completing her goal (which involves metacognitive experiences).  A theme that Flavell emphasizes throughout this brief and powerful introduction to metacognition is that the quality of people’s metacognition can impact their success and performance on a task – success is simply not just a function of the integrity of one’s cognitive system (e.g. working memory ability).   A second theme – highlighted in the article’s title – is that aspects of metacognition develop across the lifespan.  Based on this introduction to metacognition, instructors should note that: (1) metacognition is multi-faceted, (2) metacognition develops over time, and (3) educational interventions aimed at improving students’ metacognition promise to improve their achievement.

The underpinnings of metacognition pertain to how and when students rely on metacognition to guide their learning:

• Metacognitive self-regulation can occur at various stages of a students’ progress in a course: when they are studying, while they are taking a test, or after they take a test.
• Metacognitive knowledge, regulation, and monitoring comprise the foundations of self-regulated learning.
• Students use their monitoring to make decisions about how to control their learning and retrieval performance.  

The theoretical underpinnings above have a variety of implications for the role of metacognition in student achievement. There are several known benefits of metacognition:

• Students with stronger metacognitive skills have higher academic achievement, and they are more expert-like in their learning.  

• Students who are better at monitoring what they do and don’t know (i.e., are more accurate in metacognitive monitoring) can focus their studying and learning efforts on the content that they know less well. 
• Students who have better knowledge about which strategies are most beneficial to their learning are expected to be more effective learners.
• Metacognitive awareness positively correlates with problem solving skills.

But, given that metacognition is multi-faceted, each component of metacognition should be considered:

• Although accurate monitoring is essential for effective monitoring-based control, students have no direct access to how well material is represented in their mind, and hence their monitoring is based on inferences based on accessible cues that can be error-prone.
• Metacognitive knowledge, monitoring, and regulation are malleable and research has revealed that at least some of these facets can be trained through instruction.

Nelson, T. O., & Narens, L. (1990). Metamemory: A theoretical framework and new findings. The Psychology of Learning and Motivation, 26, 125-141. Nelson and Narens (1990) synthesized the growing literature on metamemory by illustrating how various isolated and seemingly fragmented lines of research were naturally inter-related.  In particular, researchers had begun to explore various monitoring judgments (e.g., judgments of learning and retrospective confidence judgments) and control processes (e.g., study time allocation and decisions to output a response during a test) but largely in isolation without fully realizing how their efforts were linked.  To integrate this research, Nelson and Narens (1990) developed a framework based on (a) the inter-play between monitoring and regulation (also called, control) that (b) occurs across the time course of learning and test taking.  The framework illustrates how these monitoring-regulation interactions can arise at any stage of learning, and at each stage, the kinds of monitoring and regulation that drive behavior may change.  Their framework was unifying in showing that some assumptions about metacognitively driven learning are the same across the learning stages, such as that accurate monitoring is important for making effective regulation decisions at study and test.  Importantly, it also highlighted that even the concept of “monitoring” and “regulation” are multifaceted, so a student can be excellent at monitoring her study progress but poor at monitoring (or evaluating) her exam performance.  Based on their framework, instructors should note that: (1) monitoring and regulation processes can occur throughout all stages of exam preparation (i.e., the metacognitive facets in the taxonomy of Figure 1 above can apply to any stage), (2) research has been conducted to explore and improve these metacognitive processes at each stage, so (3) instructors should identify which aspects of metacognition they are interested in and at what stage, which will help them use the current evidence-based teaching guide towards improving your students’ metacognition.

Metcalfe, J., & Finn, B. (2008).  Evidence that judgments of learning are causally related to study choice.  Psychonomic Bulletin and Review, 15, 174-179. A major premise for the relevance of metacognition to education is that students use their monitoring to make regulation (or control) decisions – that is, monitoring causally influences control decisions.  The authors were the first to demonstrate this causal relationship in the context of students’ decisions about choosing to-be-learned items for restudy.  In Experiment 1, college students studied word pairs (e.g. dog – spoon) during an initial trial study trial either three times or one time a piece, and on the subsequent test (e.g., dog – ?), recall performance was higher for those items that were initially studied three times.  Most important, the students studied the same word pairs on a second trial, but the number of repetitions was switched across items (with those initially receiving three repetitions only receiving 1 on the second trial, and vice versa). During this second trial, after the final repetition of an item, participants made a judgment of learning (JOL) – a measure of monitoring during study that involves predicting future recall.  They also made a control decision by deciding which items to restudy.  A restudy trial was not provided, and the second trial ended with another recall test.  The key outcome was that students’ JOLs were higher on trial 2 for items that had been studied three times on that trial than for those studied 1 time, even though recall performance did not differ for those items on the second trial.  Thus, students’ JOLs were inaccurate.  Nevertheless, students chose to restudy more of the items that received a single study trial on the second trial (vs. those receiving three).  That is, the students chose to restudy those items that they had judged as less well learned even though those items in fact had not been less-well learned.  Based on these outcomes, instructors should note that (a) students will use their monitoring about what they know (vs. don’t know) to make decisions about what to study, so (b) if their monitoring judgments are not accurate, they will make poor control decisions.  

Koriat, A., & Goldsmith, M. (1996). Monitoring and control processes in the strategic regulation of memory accuracy. Psychological Review, 103, 490-517. Koriat and Goldsmith (1996) argued that the quality and quantity of accurate test performance relies on the accuracy of evaluating the quality of answers that come to mind as one is taking an exam.  To evaluate their arguments, they introduced a research design that included two phases and conducted two experiments with undergraduates (N = 101 total) within a laboratory setting.  During the forced-report phase, participants were forced to provide answers to general knowledge questions and made a confidence judgment about the correctness of each answer.  The second phase was free report: they were asked every question again, but now they could withhold responses if they wanted.  In one experiment, the accuracy of the confidence judgments was manipulated across groups (by using different items sets that were expected to moderate accuracy), and measured confidence accuracy by correlating each participant’s judgments with his or her own performance during the first phase.  Their manipulation worked, with relative judgment accuracy being substantially greater for one group (M = .90, which is close to perfect accuracy) than the other (M = .26).  The expectations is that these differences in accuracy would substantially impact performance on the second phase, with the former group being better able to distinguish between correct answers that they should output and incorrect ones that they should keep to themselves.  And, as expected, when participants did respond with answers during the second phase, those who were initially more accurate at evaluating their answers almost always responded with correct answers during the second phase involving free report (about 75% correct) whereas those who were less accurate at evaluating their initial answers rarely responded correctly during this free-report phase (21% correct).  Based on these outcomes, instructors should note that (a) students will use their evaluations about their answers to decide how to respond, so (b) if their evaluations (i.e., confidence judgments) are not accurate, they will make poor decisions when taking tests.   

Dunlosky, J., & Rawson, K. A. (2012). Overconfidence produces underachievement: Inaccurate self-evaluations undermine students’ learning and retention. Learning and Instruction, 22, 271-280. College students studied key-term definitions from an introductory psychology textbook (e.g., definitions for proactive interference, retroactive interference, and so forth), which represent concepts that are foundational for more advanced ones.  After studying the concepts, the students were asked to try to recall each definition (e.g., “What is the meaning of proactive interference?”) and judged the correctness of their response by indicating whether it (when graded) would receive full credit,  partial credit or no credit.  After judging their response, they could restudy the correct answer for as long as they wanted.  The task was performed on a computer, which used each students’ judgments to decide when to terminate study.  For responses that a student judged as partially correct or incorrect, the definition would be retested later on, along with another judgment and restudy.  A student continued this cycle (test, judgment, and restudy) until that student judged that a response was correct on three separate test trials.  After all definitions had been dropped from practice, students received a final test (i.e., “What is the meaning of proactive interference?”) two days later.  All responses during the practice phase and the final test phase were subsequently hand scored so that the accuracy of the students’ judgments and their final test performance could be computed.  Two outcomes are noteworthy.  First, individual differences occurred in the accuracy of the judgments, with some students being quite accurate (when they indicated that a response was correct it typically was entirely correct) and others being overconfident (when they indicated a response was correct it sometimes was either partially correct or incorrect).  Second, overconfidence was inversely related to final test performance, with those being more overconfident while judging their responses during practice performing more poorly on the final test.  Based on these outcomes, instructors should realize that overconfident students may prematurely terminate their study of course content simply because they believe they know the material when in fact their confidence is misplaced.

Serra, M., & Dunlosky, J. (2010).  Metacomprehension judgments reflect the belief that diagrams improve learning from text.  Memory, 18, 698-711. Many investigations have now demonstrated a core assumption of Koriat’s (1997) cue-utilization framework of metacognitive monitoring; namely, when people assess an internal state (such as how well they have memorized a concept, how well they understand an idea, or how well they are performing on exam questions), their monitoring is based on inferences about what is being judged and does not directly tap into the internal state.  Consider outcomes from Serra and Dunlosky (2010) who had college students read a technical description about how lightning develops.  After reading each paragraph of the description, the college students were asked to make a metacomprehension judgment that entailed predicting how well they would perform on a test over the content in the paragraph.  A critical manipulation was the presence of supporting material: some participants just read the paragraphs, some read versions that also included process-oriented pictures of lightning, and others read versions with pictures of lightning strikes.  After reading, a test was administered over the content.  As expected, performance was higher when pictures were presented, but only when the pictures illustrated the processes involved in lightning strikes. By contrast, students’ judgments were higher when pictures were presented than when none were, regardless of whether the pictures improved the students’ understanding or did not.  That is, the students did not track how well the content was understood but instead were influenced by the presence/absence of pictures.  Based on these outcomes (and many other examples from the literature), instructors should realize that students may base their judgments on any number of available cues, some of which may not actually reveal the quality of what they are judging.  Such irrelevant cues abound and will bias judgments, so many students will have difficulties making accurate judgments of their learning and performance. 

Geller, J., Toftness, A. R., Armstrong, P. I., Carpenter, S. K., Manz, C. L., Coffman , C. R., & Lamm, M. H. (2017).  Study strategies and beliefs about learning as a function of academic achievement and achievement goals.  Memory, 26, 683-690. How do biology students prepare for exams?  Over 1000 college students enrolled in introductory biology courses were surveyed, and the survey included strategies that were viewed as normatively effective (e.g., self-testing and spacing practice) and less effective (e.g., allocating the majority of study time before the exam – also known as cramming).  The authors’ question was, would students’ overall GPA predict whether they adopted effective vs. less effective strategy?  As expected, students who performed better were more likely to report using self-testing and using diagrams and were less likely to report cramming for exams.  Although correlational, these outcomes (and others from other survey-based investigations) support the claim that students who know about and more frequently use better strategies also perform better.  Based on these (and other) outcomes, some students may benefit from learning about and adopting more effective approaches to learning course content and preparing for exams.

Tanner, K. D. (2012). Promoting student metacognition. CBE—Life Sciences Education, 11(2), 113-120. This resource is an essay that translates metacognition research into practical recommendations for instructors. The essay provides actionable advice that instructors can follow to promote metacognition in their students as well as in themselves.