Scholarlink

This chapter outlines steps to analyze a complex construct of interest to science education researchers in a very commonly used digital media platform, YouTube, particularly popular science education-related videos. The construct of interest is uncertainty—established as important but challenging for teachers and researchers alike to recognize and understand as many definitions and operationalizations of uncertainty as it relates to learning science exist. To study uncertainty, transcripts of videos are created using Python and the Python packages pytube and Whisper, and a two-step triangulation approach that combines a computational (a dictionary-based text analysis) and qualitative approach. In the text analysis step, transcripts of videos are searched for key uncertainty-related terms using the statistical software R. Next, qualitative coding of the transcripts is carried out, with the output from the first step as a support for the task of developing an initial set of codes for the types of uncertainty present in the science education videos.The proposed chapter contributes to the book by providing a practical guide for researchers interested in studying complex constructs using an approach that merges some of the benefits of quantitative and qualitative approaches. Python and R code are provided to support researchers to replicate and draw on the analysis carried out.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Triangulating Computational and Qualitative Methods to Measure Scientific Uncertainty

  • 章节

  • 出版模式：

    开放获取

  • 发表日期：

    2025年

查看全文
  • Joshua M. Rosenberg,
  • Hadi Bhidya,
  • Cody Pritchard

摘要

This chapter outlines steps to analyze a complex construct of interest to science education researchers in a very commonly used digital media platform, YouTube, particularly popular science education-related videos. The construct of interest is uncertainty—established as important but challenging for teachers and researchers alike to recognize and understand as many definitions and operationalizations of uncertainty as it relates to learning science exist. To study uncertainty, transcripts of videos are created using Python and the Python packages pytube and Whisper, and a two-step triangulation approach that combines a computational (a dictionary-based text analysis) and qualitative approach. In the text analysis step, transcripts of videos are searched for key uncertainty-related terms using the statistical software R. Next, qualitative coding of the transcripts is carried out, with the output from the first step as a support for the task of developing an initial set of codes for the types of uncertainty present in the science education videos.The proposed chapter contributes to the book by providing a practical guide for researchers interested in studying complex constructs using an approach that merges some of the benefits of quantitative and qualitative approaches. Python and R code are provided to support researchers to replicate and draw on the analysis carried out.