Tutoría socrática con inteligencia artificial generativa y pensamiento lógico en estudiantes universitarios de programación: Estudio cuasiexperimental
Year: 2026
Type: article
Study design: Clinical Trial
Abstract: The study aimed to compare performance on five dimensions of logical thinking between university programming students who received Socratic tutoring mediated by generative artificial intelligence and students who received regular instruction. A quasi-experimental pretest–posttest study was carried out with two intact second-year groups of the Programming III course at a regional campus of the University of Panama (experimental group, n = 10; control group, n = 8). The four-week intervention comprised five sessions with a tutoring agent deployed locally on an open language model, configured to guide students through questions and graduated support without providing solutions. A 20-item test covering decomposition, pattern recognition, abstraction, algorithmic design and applied logical reasoning was administered. The groups did not differ significantly at pretest. At posttest, the experimental group outperformed the control group in four dimensions (Holm-adjusted p ≤ .021; Hedges' g between 1.62 and 2.90). When pre–post change was compared, the advantage was concentrated in pattern recognition, abstraction and applied logical reasoning, with differences of 54 to 74 percentage points of the maximum score, whereas the difference in change for decomposition and algorithmic design was small and imprecise. Students rated the agent favourably (4.0 out of 5), and interaction logs showed fewer responses at the lowest performance level, although full autonomy was not reached. It was concluded that Socratic tutoring was associated with better results in specific dimensions of logical thinking, although the sample size, the lack of randomisation and the limited validation of the instrument restrict causal inference.
Author: Abdiel Miranda
Institution: Universidad de Panamá
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W7225990852
