Microanalytically Assessing Self-Regulated Learning During Task Performance: Conceptual Foundations and Relations With Aptitude and Event Measures

• In Event: 72.061 - Measuring Self-Regulated Learning: Issues and Innovations

Abstract

SRL microanalysis (Cleary & Zimmerman, 2001; Cleary et al., 2012) employs a structured interview format to examine an array of theoretically-linked and interdependent processes during learning. SRL microanalytic approaches are grounded in social-cognitive theoretical principles, and specifically Zimmerman’s (2000) three-phase cyclical model of SRL. As a result, this assessment methodology is designed to gather data about students’ regulatory approach prior to a task (i.e., forethought phase), their enlistment of self-control and monitoring processes during the task (performance phase), and their judgments and reactions following the task (i.e., reflection phase). Although the basic procedures and assumptions of SRL microanalysis contrasts sharply with traditional SRL assessment approaches because it elicits data about SRL “in the moment” or as the task unfolds and uses open-ended response formats to gather such information, it is important to underscore that SRL microanalytic protocols are considered a type of self-report measure because they solicit information directly from students regarding their perceptions, beliefs, and reactions.

This presentation will provide a general overview of SRL microanalytic methodology, briefly contrasting it with other common forms of SRL assessment, such as questionnaires and think aloud protocols. Particular attention will center on the core types of questions and response formats used in this approach as well as how SRL microanalytic data can serve as a useful supplement to other types of SRL assessment data when evaluating performance outcomes. Select studies from the extant SRL microanalytic literature along with a recent empirical study will be presented to: (a) illustrate the level of convergence between SRL micronalytic protocols and other event measures, such as behavioral traces, as well as more global measures, such a questionnaires and teacher ratings; and (b) explore the extent to which microanalytic data predicts unique variation in performance outcomes. In one study, 96 eighth grade students were given a microanalytic interview as they completed a set of algebraic expression problems (Callan & Cleary, 2014). Behavioral traces of the students’ strategic behaviors during mathematics problem-solving were obtained as were more global indicators of their strategic behaviors in mathematics classes, such as self-report questionnaires and teacher ratings. A key finding in this study was that the strategic microanalytic processes exhibited statistically significant relations with behavioral traces and teacher ratings but negligible and non-significant relations with student questionnaire data. The fact that the SRL microanalytic protocol correlated with one type of global SRL measure (teacher ratings) but not another (student questionnaire) has important implications for how certain SRL measures can be used collectively to make more valid inferences about students regulatory behaviors. Further, regression analysis showed that while SRL microanalaytic data and behavioral traces was more highly correlated with mathematics task performance in the study, teacher ratings were a stronger predictor of student mathematics on more long-term and distal mathematics outcomes (e.g., comprehensive mathematics exams, and standardized test scores). Subsequent analysis, however, revealed that the relations between microanalytic questions and more global, standardized outcomes may be moderated by the difficulty level of the questions that students are asked to solve.

Author

• Timothy J. Cleary, Rutgers University