Survey data from a classroom experiment on Generative AI interaction modes in object-oriented programming education

This dataset contains anonymized survey data from a classroom-based educational study comparing three Human–Generative AI interaction modes in introductory object-oriented programming tasks: No AI, AI Tutor, and Direct AI. Thirty students completed three C# OOP tasks targeting encapsulation, inheritance, and polymorphism. The survey collected baseline self-reports, task start and end times, submitted code, short explanations, perceived difficulty, completion status, conceptual/applied/reasoning/transfer answers, comparative mode perceptions, and final reflections. The survey was administered in Ukrainian; the original responses are retained in Ukrainian, while this repository provides English documentation and metadata.