AI Assistance Usage and Programming Performance: A Data Analytics–Driven Correlational Study of First-Year IT Students

Abstract

This research examined the AI assistance usage and programming performance among first-year Information Technology (IT) students at three municipal colleges in the Philippines. The study's objectives were to evaluate students' utilization of AI tools, identify usage patterns, and analyze the correlation between AI assistance usage and programming performance. A descriptive-correlational quantitative research design was used. Data were collected from 250 first-year IT students using a structured questionnaire.
The results indicate that, although students demonstrated considerable confidence in the perceived usefulness of AI (mean 3.63), their actual dependency on these tools remained relatively low (mean 2.79). Many respondents reported limited study habits, with 82% (n=205) spending only 1–3 hours per week practicing programming. The findings indicated that a significant percentage of students commenced the IT program without any prior programming knowledge or experience (79.2%), and 84% (n=210) recognized ChatGPT as their principal AI resource. Statistical analysis showed no significant differences in AI assistance usage and programming performance between male and female students (p > 0.05). Furthermore, Pearson correlation analysis revealed a statistically significant but weak positive relationship between AI assistance usage and programming performance (r = 0.1096, p = 0.001).
Multiple regression analysis further revealed that AI assistance usage alone was not a significant predictor (p = 0.182); rather, performance was primarily driven by weekly study hours (p = 0.001) and prior programming experience (p = 0.032). Crucially, a significant interaction effect (p = 0.008) confirmed that AI assistance usage effectively served as a "cognitive scaffold" for students with higher study hours, whereas high AI reliance did not yield significant gains for those with minimal practice.