教育形态的根本性跃迁也促进了教学改进方式的变革，进一步推动了由传统“基于经验”教学改进向“循证”教学改进转变。在国家政策和数字技术的双重驱动下，研究者和实践者积极尝试，探索循证改进的行动路径，开始探讨实现教学改进智能化、精准化和科学化的方法策略，近年来国际上的重要方向之一是设计分析，以此促进学习设计和学习分析领域的双向融合，从而支持循证改进。

当前研究与实践中学习分析支持的教学改进目前主要聚焦在实施态，国内外目前探索的方向是要实现从设计态到实施态的完整闭环，但是目前已有研究缺少设计态方向的有效探索，也尚未形成设计态-实施态的完整闭环。具体而言，存在以下四方面难点：1）在“设计态-实施态”的循证改进方面：关键要素和环节不明晰；2）在“设计态”的循证改进方面：操作性和计算方法模糊，很难形成自动化的教学改进路径；3）关键要素和技术瓶颈方面：包括表征、设计、解释和干预框架不明晰，学习设计表征语言尚未聚焦混合式教学情境，混合式教学的设计分析计算方法尚未厘清；4）模型应用方面：模型难以落地，尚未走向工具化。

基于以上挑战，本研究围绕“混合式教学环境下，如何从设计分析视角支持教师的循证改进”这一核心问题，以“理论模型-操作模型-应用验证”的思路开展研究。其中研究内容1和2用文献计量分析、内容分析、专家征询法解决“如何构建混合式教学的设计分析理论模型和操作模型”，研究内容3和4针对XX开放大学的混合式教学情境，采用设计性研究范式迭代验证32名教师应用混合式教学环境下的“设计态”设计分析模型的效果、进一步采用纵向对比实验验证9门混合式课程来解决“如何应用和验证混合式教学环境下的设计分析模型”。

研究一：针对设计分析研究“设计态-实施态的循证改进方面，关键要素和环节不明晰”的难点，本研究构建了设计分析的理论模型。本研究首先在已有研究的基础上界定了设计分析的概念，梳理了设计分析的层级结构，提出设计分析研究的关键难点和要素。以第四范式、闭环控制系统为理论基础，通过特征定义-理论基础-理论要素和流程的逐层演绎，构建了包含了狭义和广义设计分析，支撑不同层级设计分析研究的设计分析理论模型。

研究二：针对设计分析研究的关键难点和技术瓶颈，本研究以混合式教学为研究情境，通过文献分析和理论演绎的方式定义了操作模型构建的三个关键要素：即操作模型的表征、设计、解释和干预框架，学习设计表征语言以及设计分析算法，并逐层剖析了以上三个要素的典型特征、量化指标和计算方法，最终构建了2个关键环节和13个具体要素的混合式教学的设计分析操作模型。

研究三：针对设计态的循证改进方面操作性和计算方法模糊，很难实现自动化的教学改进路径，以及模型难以落地，尚未走向工具化的难点，本研究针对混合式教学环境下的“设计态”设计分析模型单独进行应用和验证。研究三采用设计性研究范式，通过两轮设计性迭代，应用和验证混合式教学的“设计态”设计分析模型。为保证模型的有效性，本研究设计了支持“设计态”设计分析的工具原型，并指导了工具的开发，通过教师应用“设计态”设计分析工具支持模型进一步在实践层面落地，解决设计分析目前的短板问题。研究迭代构建了四个步骤的“设计态”设计分析模型，形成了混合式教学“设计态”设计分析模型的五大设计原则，阐释了教师应用“设计态”设计分析模型的知识转换机制，明确了混合式教学“设计态”开展设计分析的原则和策略，最后从学习设计的改进、教师的混合式教学胜任力、教师对模型的感受三个方面验证了模型的有效性。

研究四：针对混合式教学环境下的设计分析模型进行应用和验证。在研究三的基础上，本研究试图应用和验证整体设计态-实施态模型，支持教师全过程证据链的混合式教学循证改进。具体而言，在改进后的混合式学习设计方案基础上，把该部分的分析结果纳入整体混合式教学环境下的设计分析模型进行应用和验证。首先，通过给9门课程提供学习设计改进建议反馈报告的方式对混合式教学环境下的设计分析模型进行应用。通过BERT语义分类模型对课程文本数据进行自动化分类，通过对不同课程的临场感水平、课程认知临场感变化趋势分布以及学习者临场感的交互模式进行综合分析，给出设计分析视角的混合式学习设计改进建议。其次，基于课程对混合式教学环境下的设计分析模型进行应用的结果，采用纵向对比实验对9门课程应用模型前后的结果进行验证。研究结果表明，混合式教学环境下的设计分析模型能够有效提升学习者的混合式学习体验、课程学习表现，同时教师对混合式教学环境下的设计分析模型持有较为积极的态度，认为模型能够为其提供学习设计和教师能力方面的支持。

本研究的成果包括理论和实践两个层面。理论成果：1）本研究构建了适用不同层级设计分析研究的系统设计分析理论模型；2）基于理论模型进一步构建了面向“设计态”和“设计态-实施态”场景可计算的混合式教学的设计分析操作模型；实践成果：1）提出了分别针对“设计态”和“设计态-实施态”设计分析的原则和策略库；2）基于模型的应用迭代过程开发了支持混合式教学环境的“设计态”设计分析的工具系统。

本研究的核心创新点包括以下三个方面。在理论方面，批判性构建了面向混合式教学环境的设计分析理论模型和操作模型；在工具方面，设计并开发了支持混合式教学“设计态”设计分析的工具系统；在方法方面，创造性提出了设计分析算法和设计分析模型验证的方法路径。本研究的成果创新不仅能够在理论层面丰富当前设计分析的内涵，在实践层面能够提供混合式教学循证改进的原则策略和方法路径，并开发了系统的技术工具平台支持设计分析落地，破解目前循证改进的关键难点，具备较好的科学与应用价值。

The fundamental transition of educational form also promoted the reform of teaching improvement mode, and further promoted the transformation from traditional “experience-based” teaching improvement to “evidence-based” teaching improvement. Driven by national policies and digital technology, researchers and practitioners actively try to explore the action path of evidence-based teaching improvement, and begin to explore methods and strategies to realize intelligent, precise, and scientific teaching improvement. In recent years, one of the important directions in the world is design analytic, which promotes the two-way integration of learning design and learning analytics, to support the improvement of evidence-based teaching.

At present, the teaching improvement supported by learning analytics in current studies and practices mainly focused on the implementation state. The current worldwide exploration direction is to achieve a complete closed loop from design state to implementation state. However, the existing studies lacked effective exploration of design state direction and have not yet formed a complete closed loop between design state and implementation state. Specifically, there were difficulties in the following four aspects: 1) in the evidence-based improvement of “design-implementation state”: the key elements and links were unclear; 2) In the “design state” aspect of evidence-based improvement: the operation and calculation methods were ambiguous, and difficult to form an automated teaching improvement path; 3) Key elements and technical bottlenecks: the framework of representation, design, interpretation, and intervention was not clear. The representative learning design language did not focus on blended teaching context, and the design analytic and calculation method of blended teaching had not been clarified; 4) Model application: The model was difficult to apply and had not yet become instrumental.

Based on the above challenges, this study focused on the core issue of “how to support teachers’ evidence-based teaching improvement from the perspective of design analytics in blended teaching context”, carrying out the study with the idea of “theoretical model-operational model-application verification”. The research contents of part 1 and part 2 used bibliometric analysis, content analysis and expert consultation methods to solve the problem of “how to build the theoretical model of design analytics and operation model in blended teaching”. Research contents of part 3 and part 4 focused on the blended teaching situation of XYZ Open University. The design research paradigm was to iteratively verify the effect of 32 teachers’ applying blended teaching of the “design state” design analytics model. The longitudinal comparison experiment was further used to verify 9 blended courses to solve the problem of “how to apply and verify the design analytics model in blended teaching”.

Study 1: Focusing on the design analytics difficulties of “key elements and transitions in which evidence-based improvement of design-implementation state were unclear”, this study constructed a theoretical model of design analytics. The study first defined the existing concept of design analytics, combed the hierarchical structure of design analytics, and put forward the key difficulties and elements of design analytics research. Based on the theoretical models of 4^th Normal Form and Closed-loop Control System, a theoretical model of design analytics, which included different narrow-broad and supportive levels of design analytics, was constructed through the deduction of feature definition - theoretical basis - theoretical elements - process.

Study 2: Focusing on the key difficulties and technical bottlenecks of design analytics research, this study took blended teaching as the research context and defined three key elements of operation model construction through literature review and theoretical deduction, which were 1) the representation, design, interpretation, and intervention frameworks of the operation model; 2) the representation language of learning design; 3) the design analytics algorithm. The typical characteristics, quantitative indicators, and calculation methods of the above three elements were analyzed layer by layer, and finally constructed the design analytics operation model in blended teaching with 2 key transitions and 13 specific elements.

Study 3: Specific to the difficulties of ambiguous operation and calculation methods for evidence-based improvement of design state, difficult to realize automated teaching improvement path, and difficult to implement the model and not yet moving toward instrumentalization, this study separately applied and verified the “design state” design analytics model in blended teaching. Study 3 adopted the design research paradigm, through two rounds of design iteration, and applied and verified the “design state” design analytics model in blended teaching. In order to ensure the effectiveness of the model, this study developed a prototype tool supporting the design analytics of “design state” and guided the development of the tool. Through the teacher-side application of “design state” design analytics tool, the model could be further implemented in practice, and the current shortcomings of design analytics can be solved. The research iteratively built a four-step “design state” design analysis model, formed five design principles of the “design state” design analytics model in blended teaching, explained the teacher-side knowledge transfer mechanism to apply the “design state” design analytics model, and defined the principles and strategies for evidence-based improvement of “design state” in blended teaching. Finally, the effectiveness of the model was verified from three aspects: the improvement of learning design, the competence of teachers in blended teaching and the teachers’ feelings of the model.

Study 4: Application and validation of the design analytics model in blended teaching. On the basis of Study 3, this study attempted to apply and verify the overall design-implementation model to support the evidence-based improvement in blended teaching of the whole process evidence chain for teachers. Specifically, on the basis of the improved blended teaching design scheme, these analytic results were incorporated into the overall blended teaching design analytics model for application and verification. Firstly, the design analytics model in blended teaching was applied by providing feedback reports on the improvement suggestions of learning design for 9 courses. By using BERT semantic classification model to automatedly classified the course text data, and through a comprehensive analysis of the presence level of different courses, the distribution of changing trend in curriculum cognitive presence, and the interaction mode of learners’ presence, the suggestions for improving blended teaching design from the perspective of design analytics were put forward. Secondly, based on the results of the application of design analytics model in blended teaching, the longitudinal comparison experiment was used to verify the results before and after the application of the model in 9 courses. The research results showed that the design analytics model in blended teaching could effectively improve learners’ blended learning experience and course academic performance. Meanwhile, the teachers held a positive attitude towards the design analytics model in blended teaching, believing that the model could provide support for learning design and teachers’ ability.

The results of this study included theoretical and practical aspects. Theoretical results: 1) This study built a systematic design analytics model applicable to different levels of design analytics research; 2) Based on the theoretical model, the calculatable design analytics and operation model in blended teaching facing “design state” and “design-implementation state” scenarios were further constructed; Practical results: 1) The principles and strategies of evidence-based improvement for “design state” and “design-implementation state” were proposed; 2) A model-based application iterative process was developed to support the evidence-based improvement of the “design state” systematic tool in blended teaching.

The core innovation of this study included the following three aspects. In the aspect of theory, the theoretical model of design analytics and operation model in blended teaching scenario were critically constructed. In the aspect of tools, a systematic tool supporting the evidence-based improvement of the “design state” in blended teaching was designed and developed. In the aspect of method, the method path of design analytics algorithm and model verification of design analytics was innovatively proposed. The innovation results of this research could not only theoretically enrich the meaning of current design analytics, but also practically provide principles, strategies and methods for evidence-based improvement in blended teaching. A systematic platform of technical tool to support the implementation of design analytics was developed, so as to solve the key difficulties of the current evidence-based teaching improvement, which had scientific and application values.