Exploring Data Science Students' Engagement, Usage Patterns, and Perceptions of Large Language Models in Programming.

Area of Science:

• Artificial Intelligence
• Educational Technology
• Data Science

Background:

• Large Language Models (LLMs) are increasingly adopted by students for academic tasks.
• Understanding student perceptions and usage patterns of LLMs in programming is crucial.

Purpose of the Study:

• To investigate how Master's students in Data Science for Health perceive and utilize LLMs for programming assignments.
• To identify common LLM applications, usage frequency, and student satisfaction levels.

Main Methods:

• A survey was conducted among 77 students and recent graduates of a Master's in Data Science for Health program.
• The survey focused on their experiences with LLMs for programming tasks like debugging, code generation, and error analysis.

Main Results:

• 84.4% of respondents use LLMs for programming; 55.4% use them daily.
• Users reported satisfaction with LLM speed, ease of use, and debugging, but raised concerns about dependency, data privacy, and reference accuracy.
• 87.7% actively monitor LLM developments, and most verify LLM-generated outputs.

Conclusions:

• LLM integration into educational settings is essential for promoting responsible use and best practices.
• Educational strategies and assessments need adaptation to incorporate LLMs while emphasizing fundamental skills like problem-solving and critical thinking.