Ayush Pandey

Student self-efficacy in computing at the high school level

Many existing AI-focused educational programs for high school students are basic primers and lack technical depth. Can we design curricula such that high school students can learn the technical concepts underlying neural network design? A hypothesis is that research project scaffolds may be helpful towards that end. Another question is to study the effect of various kinds of teacher-scholar-mentor models on student self-efficacy in computing. For example, does a communication TA along with a "regular" TA enhance the educational output? What about learning assistants? There are two elements in this research project: experiments with high school programs, and data analysis of existing programs. For the first line of research, we need to devise strategies and collaborate with AI programs for high school students. For the second line of research, we need to develop mixed-methods for data analysis (student feedback, course evaluations, student reflections, etc.)

Research questions:

  1. Does a pre-college program featuring directed research and communication mentoring enhance students' comprehension of neural networks, confidence, and readiness for a college major in computer science or related engineering disciplines?
  2. Which specific aspects of self-efficacy and social and emotional learning are most affected among students who participated in a rigorous, AI-focused technical summer program
  3. What is the correlation between pre-program responses on "career goals" with the post-program reflections after two years on career readiness?
  4. Methods research: What methods can validate heuristic thematic analysis with automated approaches for a reliable study of student reflections from qualitative data?
  5. Methods research: What is the accuracy of LLMs in quantifying themes as compared with conventional thematic analysis process? Specifically, can we quantify student comprehension of neural networks, their outlook on AI, and appreciation for AI applications among students using their subjective survey responses?

Participants

  1. Thomas Williams, UC Merced

Collaborators

  1. S. Shailja, Stanford University

Past participants

  1. Satish Yadav, UCSB
  2. Arthur Caetano, UCSB

Publication

  1. S. Shailja, S. Yadav, A. Caetano, A. Pandey, "Scaffolding AI research projects increases self-efficacy of high school students in learning neural networks (Fundamental)" (accepted) American Society of Engineering Education (ASEE) Annual Conference and Exposition, 2024.

Last update: Mar 15, 2024. You can contribute to this page by creating a pull request on GitHub.