Treffer: Using Hierarchical Multiple Regression to Model the Impact of AI-Powered Adaptive Testing on Student Academic Fortunes and Test Anxiety in the Ghanaian Context

This study investigated the impact of AI-powered adaptive testing on student academic performance and test anxiety, comparing its effectiveness to traditional testing methods. Using a quantitative research approach, hierarchical regression analysis was employed to examine the influence of adaptive testing on student outcomes, controlling for variables such as demographics, prior academic performance, and technological familiarity. A sample of 250 senior high school students in Ghana was selected via stratified random sampling to ensure diversity. Results indicated that AI-powered adaptive testing significantly improved academic performance, with a mean score increase of 7.3 points over traditional methods (t = 5.32, p < 0.001), and reduced test anxiety levels by a mean difference of 3.8 points (t = 4.87, p < 0.001). Hierarchical regression analysis showed that adaptive testing accounted for an additional 20% of the variance in academic performance and 16% in test anxiety, underscoring its effectiveness in supporting student outcomes. Additionally, student demographics, prior academic performance, and technological familiarity were significant predictors of adaptive testing's effectiveness. Normality tests confirmed the robustness of these regression results. Findings suggest that AI-powered adaptive testing provides a more tailored, effective approach to student assessment, potentially enhancing educational outcomes and reducing anxiety. Recommendations include integrating adaptive testing systems into educational practices, training educators, addressing technology access disparities, and conducting further research to assess long-term impacts across diverse educational settings.

As Provided