SPSS Analysis: GenAI Use, Writing Self-Efficacy, & Trust

 This study analyses the relationship between GenAI use, trust, and writing self-efficacy among university students. 102 survey responses were collected, then analysed via SPSS.

Findings & Implications

1. The correlation between GenAI use and writing self-efficacy was positive but not statistically significant; r(100) = .10, p = .34.
2. There was a significant association between GenAI use and trust in GenAI; r(100) = .53, p < 0.01.
3. There was no significant relationship between trust in GenAI and writing self-efficacy; r(100) = -.02, p = .83. 
4. A moderation analysis was conducted to examine whether trust in GenAI moderates the relationship between GenAI use and writing self-efficacy, controlling for gender, age, and level of study. The overall model was not significant, F(6, 95) = 1.20, p = .31, with an R² of .07. The interaction term between GenAI use and trust in GenAI was not significant, β = .05, t(95) = 1.43, p = .16, 95% CI [-.02, .12].
5. A one-way ANOVA was conducted to examine whether students at different levels of study (i.e. Bachelor’s, Master’s, PhD) differed in their use of GenAI for academic writing, trust in GenAI, and writing self-efficacy. The results showed no significant differences across study levels for trust in GenAI, F(2, 99) = 0.46, p = .632, GenAI use, F(2, 99) = 0.38, p = .682, or writing self-efficacy, F(2, 99) = 0.41, p = .665.
6. A one-way ANOVA was also used to test for differences among gender groups. The effect of gender on GenAI use was marginally non-significant, F(3, 98) = 2.65, p = .053; no significant gender differences emerged for writing self-efficacy, F(3, 98) = 1.89, p = .136, or trust in GenAI, F(3, 98) = 1.14, p = .338.
   • 

While frequent GenAI use correlates with higher trust, neither GenAI use nor trust significantly influences writing self-efficacy, and trust does not moderate the effect. Hence, there is a need to explore how writing self-efficacy adapts in rapidly evolving, technology-rich learning environments, and how human agency interacts with AI mediation in shaping academic outcomes. Other factors worth investigating are GenAI reflective experiences, mastery, purpose of use — they are likely to influence writing self-efficacy as well.

Full Paper & Dataset