Developing the temporal analysis for computer-supported collaborative learning in the context of scaffolded inquiry

Abstract

Computer-supported collaborative learning (CSCL) frequently takes the form of inquiry-based learning (IBL) in science education. To achieve the benefits of computer-supported collaborative inquiry-based learning (CSCIL), various scaffolds have been studied from the perspective of what (not how) learning occurs and what (not how) differences emerge between the scaffolded and non-scaffolded conditions. To better address the how questions, my theoretical aim was to develop a temporal analysis procedure for CSCL. Based on a systematic literature review of 78 journal papers, I defined six key operations for the analysis of CSCL’s temporal aspects: proposing research aims regarding the temporal aspects, setting up the context, collecting process data, conceptualising events, conducting temporal analysis methods and interpreting the outcomes. A study of how the included papers performed these operations showed how the researchers implicitly conceptualised the temporal aspects of CSCL when focusing on the characteristics of or interrelations between events over time. My methodological aim was to advance temporal analysis methods to study CSCIL. My empirical aim was to design scaffolds and analyse their role in CSCIL by employing the key operations and advanced methods when groups used a numerical problem-solving tool (Python program) to inquire in undergraduate physics courses. To study how CSCIL occurs, I used video data and visualised the transitions between the IBL phases (i.e. IBL sequences) and groups’ ways of using the Python program for inquiry over time (two groups, n = 10). The identified challenges and productive practices guided the scaffold design. To study how differences emerge between the conditions (46 groups, N = 231), I performed temporal log data analysis (TLDA) and temporal lag sequential analysis (TLSA). Temporal distinctions in how the groups used the Python program between the conditions (captured by TLDA) were associated with the differences in the content and temporal emergence of IBL sequence clusters between the conditions (captured by TLSA of video data). This dissertation demonstrates how temporal analysis may advance our understanding of the premises for successful learning and benefit the design and implementation of scaffolds.