Advancing Computational Grounded Theory for Audiovisual Data from STEM Classrooms

该提案是针对电子病历核心研究(ECR)计划公告NSF 19-508而提交的。ECR STEM教育基础研究计划为关键的研究领域提供资金，这些领域是基本的、广泛的和持久的。EHR寻求的建议将有助于综合、建立和/或扩大以下重点领域的研究基础：STEM学习、STEM学习环境、STEM劳动力发展和扩大STEM的参与。ECR计划的特点是强调积累可靠的证据，为以下工作提供信息：(A)理解，(B)建立解释理论，(C)建议干预措施(和创新)，以解决STEM兴趣、教育、学习和参与方面的持续挑战。这个EHR核心研究项目正在对视频数据的计算分析进行方法论研究，重点是课堂上STEM学习的社会和空间维度。视频数据是复杂的。它们涉及视觉、听觉、空间和时间特征，这些特征可以通过几种方式减少。到目前为止，对STEM教室的视频数据的分析还无法利用计算能力来利用其丰富性。然而，数据科学的最新进展，加上现有的语音分析方法，使得以保持复杂性和细微差别的方式从视频中通过计算识别重要特征成为可能。这些进展将提高研究的可复制性。通过该项目开发的方法将促进更多的研究人员对视频数据进行复杂的计算分析。这些新方法的应用将有助于增加视频研究的规模和普适性，并导致新理论的建立。这项研究项目建立在最先进的计算机视觉和语音分析方法的基础上，这些方法是在STEM教室收集的视频数据上进行测试的。它是在一个基于计算的理论方法论框架内做到这一点的，该框架利用了扎根分析方法的解释能力和计算方法的处理能力。具体来说，将产生两种类型的计算分析程序：(A)从STEM教室的视频和音频数据中提取有意义的特征，以及(B)使用这些提取的特征进行探索性模式识别。为了制定这些程序，将使用现有的STEM教室的大规模视频数据集来测试和改进日益复杂的分析，这些分析还将用于演示这些方法的应用，以调查STEM教室的社会和空间维度。该项目的重点是整合这些方法，以提高它们的能力，并利用STEM教室现有的大规模数据集，以便开发的方法可以在现实数据上进行测试。这些数据集足够广泛，足以支持对广泛的研究问题的调查，包括关于学生参与纪律实践的高推断性问题。最后，通过配对计算和基础分析方法，该项目正在开发有潜力增强和测试数据中发现的模式的结构有效性的方法。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This proposal was submitted in response to EHR Core Research (ECR) program announcement NSF 19-508. The ECR program of fundamental research in STEM education provides funding in critical research areas that are essential, broad and enduring. EHR seeks proposals that will help synthesize, build and/or expand research foundations in the following focal areas: STEM learning, STEM learning environments, STEM workforce development, and broadening participation in STEM. The ECR program is distinguished by its emphasis on the accumulation of robust evidence to inform efforts to (a) understand, (b) build theory to explain, and (c) suggest interventions (and innovations) to address persistent challenges in STEM interest, education, learning, and participation.This EHR Core Research project is conducting methodological research on the computational analysis of video data focused on the social and spatial dimensions of STEM learning in classrooms. Video data are complex. They involve visual, acoustic, spatial, and temporal features that can be reduced in several ways. To date, analysis of video data of STEM classrooms has not been able to leverage computational power to take advantage of their richness. However, recent advancements in data science, coupled with existing speech analytics methods, make it possible to computationally identify important features from video in ways that preserve complexity and nuance. These advancements will improve research replicability. The methods developed through this project will facilitate use of sophisticated computational analysis with video data by more researchers. Application of these new methods will help increase the scale and generalizability of video research and lead to the building of new theory. This research project builds on state-of-the-art computer vision and speech analytics methods tested on video data collected in STEM classrooms. It does so within a computational grounded theory methodological framework, which leverages the interpretive power of grounded analytical approaches with the processing power of computational methods. Specifically, two types of computational analysis procedures will be produced: (a) extracting meaningful features from video and audio data of STEM classrooms, and (b) conducting exploratory pattern identification using these extracted features. To develop these procedures, existing large-scale video datasets of STEM classrooms will be used to test and refine increasingly sophisticated analyses, which will also be used to demonstrate the application of these methods to investigate the social and spatial dimensions of STEM classrooms. The project focuses on integrating these methods to improve their power and leverages existing large-scale datasets of STEM classrooms, such that the methods developed can be tested on realistic data. The datasets are extensive enough to support the investigation of a wide range of research questions, including high-inference questions about students' participation in disciplinary practices. Finally, by pairing computational and grounded analytical methods, the project is developing methods that have the potential to enhance and test construct validity of the patterns found in the data.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.