Aims

            Of late, there is growing concern about the clarity of meaning with regard to three often used and seemingly related terms, metacognition, self-regulation (SR), and self-regulated learning (SRL). It seemed from our cursory look into the literature that the rising popularity of self-regulation and self-regulated learning and the reemerging interest in metacognition added to this dilemma. Therefore, we chose to look deeper into the literatures on metacognition, self-regulation, and self-regulated learning. We consider the disentangling of these terms as an important step to advance programs of research on  monitoring and regulation of self and others

            Our focus for this review was to explicate the core meaning of metacognition, self-regulation, and self-regulated learning, as well as determine where these constructs converge and diverge along select dimensions. To accomplish these goals, it was essential to attend not just to the recent research in these areas, but to the theoretical and empirical roots of each. By tracing back to the empirical beginnings of metacognition, self-regulation, and self-regulated learning, we felt we could better understand how and why the conceptualization and operationalization of these terms evolved into their present-day manifestations.

Methodology/Research Design

            In order to consider the state of metacognition, SR, and SRL in contemporary educational psychology literature, we created three study pools corresponding to the three constructs. These pools were developed by searching PsychInfo using the keywords, metacognition, self-regulation, and self-regulated learning, in separate searches. The keywords were not limited to title or abstract, in order to include all potentially relevant studies. Due to the vast number of studies initially identified s, we subsequently limited the search to empirical works published in the past five years. We conducted an additional hand search and limited our studies to those focusing on academic learning. The resulting pool included 255 studies (i.e., 123 metacognition, 54 self-regulation, and 78 SRL articles).

            We organized studies into a comprehensive table with three designated sections corresponding to the constructs. We analyzed each article according to the following: definitions of the constructs; measures used; measurement method (e.g., self-report); and instances of priming or cueing during a task. Analyzing the studies in this manner proved difficult for three reasons: (a) the extent to which each study identified through our search parameters dealt expressly with any of the three constructs of interest proved problematic, (b) many of the studies identified by the search parameters in PsychInfo referred to a different construct than the one keyed in the database, and (c) some studies defined and measured more than one construct.

            Due to the large number of studies and the difficulty of differentiating between them for certain dimensions, we felt that an explicit scheme was essential to ensure that resulting codes were reliable and valid. Thus, we developed two broad categories of codes for each construct: definitions and measurement. Within the category definitions, we identified the extent to which authors explicitly defined the target constructs and identified trends in definitional emphasis. For the category measurement, we coded data sources (e.g., self-report or interview) and the alignment of definition to measurement; that is, the correspondence between conceptual and operational definition. Lastly, we examined studies to determine the level of prompting or cueing in the experimental task or environment. For the table and the coding, we engaged in extensive discussion to ensure that codes were consistent. In addition, initial codes underwent modification early in the process to accommodate the diversity we encountered in the literature. Once issues with the coding scheme were resolved, the first and third authors classified all identified studies with a high level of interrater agreement (α > .85).

Findings         

            The four main findings of this review included patterns related to: the explication of definitions, definition and measure alignment, types of measures, and prompting/cuing. Regarding the explication of definitions, we found that  only 49% of all studies explicitly defined the construct under investigation. Between the three constructs, self-regulated learning and self-regulation had a much higher percentage of explicit definitions than metacognition. Three possible explanations for this pattern may be that: (a) SR and SRL have closer ties to neobehaviorism, where the requirement to provide explicit operational definitions is more common, (b) perhaps researchers assume metacognition is a rather well-established term that needs no further clarification beyond a suitable citation, or (c) it is also possible that there is more theoretical diversity represented in the SR and SRL research as compared to the more clearly cognitive perspective signaled by metacognition.

            We found that the measure alignment, measure type, and scaffolding/cuing were closely related. Overall, data for definition-measure alignment suggest that alignment was better for the SRL literature than for metacognition and self-regulation. However, these data and the coding scheme employed may not tell the complete story. We also found issues of measurement clarity, scope, type of measurement, and scaffolding. The types of measures used in the tabled studies varied widely. From our examination of these data, it was clear that self-report was the dominant type of measure for all three constructs, at nearly 43%. However, different patterns emerged when the constructs were examined separately. The most obvious trend in the data with regards to prompting/cuing was the limited presence of scaffolding in self-regulated studies. This outcome likely reflects the infrequent reliance on experimental or performance tasks in these studies. However, for metacognition and SRL, which both had a greater number of studies involving experimental or performance tasks, important differences with regards to prompting were evident.

Theoretical and Educational Significance