Investigating Students' Attitudes towards Artificial Intelligence in Somalia - Figure 2

Kruskal-Wallis test results indicated no statistically significant differences in AI Attitude scores between age groups. However, significant differences were found between age groups when AI Experience was statistically controlled using a nonparametric Quade ANCOVA. This suggests that AI Experience was a confounding variable masking the effect of age in the original analysis. By controlling for AI Experience, we revealed the independent effect of age on AI Attitude. In conclusion, these findings emphasise that variables such as AI Experience should definitely be controlled when assessing AI attitudes. Controlling for covariates reveals hidden effects and provides a more accurate understanding of differences between groups.     Although initial Kruskal-Wallis analysis suggested that the youngest group (18–20) had slightly higher AI attitude scores, these differences were not statistically significant. However, once AI experience was controlled for in the Quade ANCOVA, older students (24–26 and 27+) exhibited significantly higher adjusted AI attitude scores. This reveals that AI experience was a confounding variable, masking the true age-related effect. See full paper here: brain.edusoft.ro