Metacognition Diagnosis Research

Metacognition,as a fundamental and essential literacy ability for the general domain,is the key ability that supports learners in solving complex problems in real situations and helps them adapt to the rapidly changing and complex social environment.Therefore,accurate,comprehensive,and appropriate metacognitive diagnosis is a prerequisite for promoting learners’ metacognitive development.However,due to the complexity,multi-level,and implicit nature of metacognition,traditional metacognitive diagnostic methods are difficult to accurately quantify the complex and dynamic metacognitive activity process.Although learning analytics technology born in the era of big data provides a technical foundation for measuring and diagnosing dynamic metacognitive activities in real learning activities,it still faces problems such as data-driven,causal confusion,and lack of validity.Against this background,this study proposes to improve the limitations of current research in the field of learning analytics through a design-driven learning analytics paradigm for "metacognitive diagnosis based on learning analytics" in order to achieve a comprehensive,systematic,and dynamic diagnostic of metacognition.In order to achieve this research goal,this study conducted relevant research in the following four aspects.(1)To build a design-driven learning analytics paradigm,this study drew on the Kuhn paradigm theory in philosophy of science and the historical experience of paradigm shifts in psychometrics to analyze the problems currently existing in the domain of learning analytics.Base on this,the key to breaking through the limitations of current learning analytics was identified as "design".By designing model of the theory of learning,model of rules for evidence,model of environmental features,and model of learning analytics,a learning analytics paradigm rooted in learning science theory,consistent with the underlying logic of psychometrics,and producing analyzable results with explanatory significance was summarized.(2)Using the design-driven learning analytics paradigm,this study organizes,relates,and reconstructs the learning science theory and learning technology design methods related to metacognition,focusing on the design of four models: the learning theory model,evidence rule model,environment feature model,and learning analytics model.Based on this,a framework for metacognitive diagnosis based on learning analytics is designed from a theoretical perspective.(3)Based on the metacognitive diagnosis framework designed at the theoretical level,this study uses front-end and back-end development technologies such as Java Script,React,and Node.js to design and develop a web-based metacognitive diagnosis system.(4)The metacognitive diagnosis system designed in this study will be tested in practice.By analyzing and diagnosing the learning activity data of participants in the diagnosis system,the reliability and validity of the metacognitive diagnosis based on learning analytics will be verified.Based on the experimental results obtained using the metacognitive diagnostic system,it shows that this system has the ability to reveal the implicit metacognitive processes.Furthermore,metacognitive diagnosis based on the design-driven learning analytics paradigm can achieve systematic and processoriented diagnosis through quantitative and visual analysis.This approach can analyze and diagnose learners’ metacognitive activity states,the interactions between different metacognitive elements,and the level of metacognitive system ability.Specifically,based on this approach,it is possible to uncover the "zone of proximal development" for learners with different metacognitive abilities,which refers to the learners’ potential for metacognitive development.This approach can provide individual and interpretable metacognitive development recommendations for specific learners.