Bots, maths, and self-determination: Gendered profiles of motivation, competence, value, autonomy, and pressure in secondary school

Abstract

: The rapid diffusion of AI chatbots into classrooms has outpaced evidence on gender-specific engagement with these tools. This study compares male and female motivational chatbot-usage profiles among upper secondary mathematics students to inform gender-sensitive design and implementation. Austrian students ( N = 1,253; 46% female; ages 14–19) answered questionnaire items indexing five constructs: enjoyment, perceived pressure, choice, competence, and value. A confirmatory factor analysis supported an acceptable measurement model, and multi-group testing confirmed metric and scalar invariance across gender, supporting valid mean-level comparisons. Three profiles emerged for males and six for females. Males were dominated by a cautious-pragmatist class with mildly positive appraisals and low strain, alongside a high-adopter class and a disengaged class. Females showed a more differentiated structure, including a high-adopter class, a pragmatic positive class, an engaged but pressured class, a sceptical low-intention class, a disengaged class, and a distinct high-pressure anxious rejecter class. Grouping profiles into low/mid/high tiers for cross-gender comparability, tier prevalence differed by gender (omnibus χ 2 (2) = 22.6, p <.001): male students were roughly two times more likely than female students to fall in the high-adopter tier (18.6% vs 9.4%, p <.001). Classification-error-corrected tests showed that background variables did not differ significantly across the three male profiles; for females, age differed significantly across the six profiles ( p = .004), with weaker trends for parental education and aspiration. Findings indicate a need to scaffold competence and signal value for all students, with targeted pressure reduction particularly relevant for the female anxious-rejecter profile.