In 1976 Flavell coined the term metacognition: “Metacognition refers to one’s knowledge concerning one’s own cognitive processes and products or anything related to them. […] Metacognition refers, among other things, to the active monitoring and consequent regulation and orchestration of these processes in relation to the cognitive objects on which they bear, usually in the serve of some concrete goal or objective” (Flavell, 1979).

Twenty years after Flavell gave a name to the ball that was rolling across disciplines, the search for a higher level of modelling cognition was still being pursued. “Traditional developmental research in memory and reasoning, as well as current investigations in such disparate areas as theory of mind, epistemological understanding, knowledge acquisition, and problem solving, share the need to invoke a meta-level of cognition in explaining their respective phenomena.”(Kuhn, 2000). At that time there were several models and definitions of metacognition. These are discussed, summarized and a usable definition is distilled in 2002 by Pintrich:

“Metacognitive knowledge includes knowledge of general strategies that might be used for different tasks, knowledge of the conditions under which these strategies might be used, knowledge of the extent to which the strategies are effective, and knowledge of self ”.

Three types of metacognitive knowledge follow the categorisation set out by Flavell in 1979 (note that metacognitive control is left behind in this discussion):

Strategic metacognitive knowledge
Task oriented metacognitive knowledge
Person oriented metacognitive knowledge

”Strategic knowledge is knowledge of general strategies for learning, thinking, and problem solving.” (Pintrich, 2002). The term ‘general’ here means that strategic knowledge only concerns those strategies that are not specific to one domain or type of content for problem solving. Kuhn proposed the term ‘metastrategic knowledge’ to refer to knowledge on the meta-level of procedural knowing (Kuhn, 2000). She makes a similar distinction between metacognition on the strategic and task level but applies a hierarchical order in the distinction by first splitting metacognitive knowledge into two categories based on the type of knowledge that the meta-level refers to – procedural versus declarative knowing- and then splitting it into roughly the same ‘strategy’ knowledge versus ‘task ’ oriented knowledge.

A knowledge taxonomy

By applying taxonomy to explore the concept of knowledge, an organisational structure can be superimposed. Such a taxonomy would provide a shared structure and vocabulary for a complicated concept that cuts across many scientific fields and models.

Blooms’ original Taxonomy was devised to standardize test items for measuring educational objectives. It has been translated into 22 languages and widely used. The original Taxonomy was often used to analyse a curriculum on whether all categories were being sufficiently taught. Such analysis tended to show that there was a large emphasis on recognition and recall and not enough on items from the categories Comprehension and Synthesis – which are widely regarded to be vital goals of education.

The change we see in going from the Original to the Revised Taxonomy is the emphasis on knowledge of and responsibility over ones’ own cognition. “This change cuts across all the different theoretical approaches to learning and development-from neoPiagetian models, to cognitive science and information processing models, to Vygotskian and cultural or situated learning models.” Research finds that students learn better as they act on this awareness of their own thinking. Different fields approach this development in different ways and make mention of metacognitive knowledge, metacognitive awareness, self-awareness, self-reflection, and self-regulation. (Pintrich, 2002)

The revised Taxonomy has adapted itself to the thinking and wording of the developing field of cognitive psychology. It added a new subcategory to the Knowledge dimension: Metacognitive Knowledge. The revised Taxonomy has also discerned two dimensions: The Knowledge dimension and the Cognitive Processing dimension. We can apply the revised Taxonomy to look at any learning outcome and judge whether – or to what degree – metacognition was involved. It is a structure that can help locate Metacognition. The categories go up in complexity and form a hierarchical scale.

Knowledge on cognitive tasks and the difference between them is needed in order to prepare accordingly. Conditional knowledge depends on local situations and on more general social situations, on conventions that may or may not apply and prevalent cultural norms. For all of these conditions different strategies might be suitable. Students should know the ‘What’ and ‘How’ as well as the ‘When’ and ‘Why’ of the different learning strategies and use them appropriately (Paris, Lipson, & Wixson, 1983).

The process of achieving and adapting metacognitive knowledge can be seen as a feedback cycle. From the meta-level stem the choices of what strategies are applied.

“The meta-level directs the application of strategies, but feedback from this application is directed back to the meta-level. This feedback leads to enhanced meta-level awareness of the goal and the extent to which it is being met by different strategies, as well as enhanced awareness and understanding of the strategies themselves, including their power and limitations. These enhancements at the meta-level lead to revised strategy selection. These changes in strategy usage in turn feed back to further enhance understanding at the meta-level, in a continuous cycle in which the meta-level both directs and is modified by the performance level.”(Kuhn, 2000).

This corresponds with findings that adults are generally better at metacognition than children. “Developmentally, then, increasing meta-level awareness and control may be the most important dimension in terms of which we see change (Kuhn, 2000).

Measuring metacognition

Semerari and his colleagues – taking the perspective of the clinician – wanted to divide metacognition in several sub-functions to determine, measure and work with it in a therapeutic setting (Semerari et al., 2003). They first developed the Metacognition Assessment Scale to investigate individual verbalization in psychotherapy transcripts. Based on clinical literature concerning disorders in the ability to know and to regulate mental states, they discerned three sub-functions:

Understanding one’s own mind
Understanding others’ minds’
Mastery’.

This approach prefers the definition of Wells and Purdon (1999) of metacognition “the aspect of information processing that monitors, interprets, evaluates and regulates the contents and processes of its organization” and often refers to metacognition as a skill. This Metacognition Assesment Scale (MAS) has shown acceptable levels of factorial validity, inter-rater agreement, internal validity and test–retest stability as well as demonstrating a connection to executive functions and treatment outcomes (Semerari et al., 2012). Building on the experiences with the MAS, Semerari and colleagues developed the Metacognition Assessment Interview (MAI) which evaluates metacognitive skill in the Self-domain and the Other-domain. For the Self, monitoring and integrating was measured as a metacognitive skill while for the Other differentiating and decentring was measured. Each of these four dimension was built on four facets, making a total of 16 basic facets.

Going from the MAS to the MAI “the authors took into account the clinical literature that describes deficit in the ability to know and regulate mental states, theoretically based on the literature on mentalization and attachment theories, theory of mind, metacognition and, more generally, metarepresentation.” (Semerari, 2012).

Based on a theoretical framework where metacognition has two prerequisite parts ( notion of self and notion of other) and these three (self/other/meta) can be detected separately – research done by analysing transcript of therapeutic sessions along with a three-dimensional questionnaire. One underlying construct is found to connect all three parts – this would be metacognition influencing all the objects in the model – however separation between the concepts cannot fully be established. Further research necessary but encouraging results – use method with caution and check for updates! Also get the actual survey and method of analysing the sessions.

Veenman (1993) developed an approach to measuring the use of metacognitive skills that is based on the systematic approach to problem solving of Mettes and Pilot (1980). This approach consists of a hierarchical set of problem-solving activities that is compiled at the task-level. Veenman’s starting point are the following metacognitive strategies:

Reflecting on the nature of a problem
Comprehension monitoring
Predicting the consequences of an action or event
Planning of activities
Monitoring the ongoing activities
Testing for plausibility
Reflecting on one’s learning performance

Learning

“Metacognitive learning is assumed to play an important role in learning, specifically in the constructivist paradigm. In the constructivist paradigm, reflection and self-regulation is much more important than in earlier learning paradigms. Self-regulated learning, for instance, encompasses metacognitive, motivational and behavioural aspects of learning.” (Christoph, 2006.)

Learning strategies
“Although there are a large number of different learning strategies, they can be grouped into three general categories: rehearsal, elaboration, and organizational (Weinstein & Mayer, 1986).” Rehearsal is the much used ‘rinse and repeat’ approach of content repetition. Elaboration might include the use of mnemonics, paraphrasing or summarizing content. It has been shown that such elaboration strategies lead to a deeper level of processing and better comprehension of the content as compared to rehearsal strategies (Pintrich, 2002). To apply organizational strategies would be to connect certain elements of the content by note taking or, for example, creating a mind-map. Metacognition is applied in the planning, monitoring and regulating the use of various learning strategies. Such metacognitive knowledge concerning the different learning strategies and their (conditional) application seems to be involved in the degree to which transfer of learning occurs. Transfer or Transference is the ability to use knowledge learned in one setting or situation in another setting or situation (Bransford & Schwartz, 1999).

Interactive metacognition
A distinction can be made between this self-directed metacognition and the other-directed metacognition. Schwartz and his colleagues (2009) term the other-directed metacognition as “interactive metacognition”. Learning-by-teaching has been found to be an effective way to learn, as long as the teacher (student) remains engaged with the student and attentive to their understanding of the content and their learning process. If the teacher (student) loses the engagement and the teaching becomes a monologue, the teacher (student) learns less from this learning-by-teaching compared to self-study (Schwartz et al., 2009).

The demand of metacognition on working memory is two-fold (Schwartz et al., 2009):

The problem solving thoughts
Monitoring and regulating the thinking about the problem solving thoughts

Where one can take up resources for the other in such a way that when the problem solving thoughts are unfamiliar or difficult i.e. still in the process of learning, this drains the resources for monitoring and regulating capacity. In a teacher-student situation the distribution of resources might be shared as the teacher offers cognitive resources for the actual problem solving or for the monitoring and regulating. This sharing of working memory resources is an application of distributed cognition (Schwartz et al., 2009).

Learning and performance OR acquiring and outcome
Two processes are going on that are often viewed as the same but need to be separated are learning and performance. Learning is the process of acquiring new knowledge, attitudes or skills regardless of whether they are ever actually used. The acquired knowledge might be internalized and possibly locked away never to be seen again, it remains learned. Another process takes place when such learned content finds its way outside and is displayed, applied or otherwise shared. The learned content is then performed (Buckley & Anderson, 2006).