AI Insight
This study analyzed mathematics attitudes among 4,820 eighth-grade Turkish students and identified five distinct profiles ranging from low overall attitudes with very low confidence to highly positive attitudes across all dimensions. School belonging was the strongest predictor of attitude profile membership, followed by home educational resources, while bullying victimization showed no independent association when school belonging was included in the model. Gender showed only a modest relationship with mathematics attitude profiles.
Why it matters
The findings suggest that fostering students' sense of school belonging may be more important for developing positive mathematics attitudes than directly addressing bullying, though the two factors appear interconnected. Schools could potentially improve mathematics engagement by prioritizing interventions that strengthen students' connection to their school community and by considering home educational contexts.
Understand the Science
IntroductionThis study examined heterogeneity in students’ mathematics attitudes and the associations of perceived school belonging, bullying victimization, gender, and home educational resources with profile membership.MethodsTIMSS 2023 Türkiye data included 4,925 eighth-grade students (LPA analytic N = 4,820; three-step analytic N = 4,739). Latent profile analysis used the three official TIMSS derived scale scores for liking mathematics, confidence in mathematics, and valuing mathematics, with equal indicator variances across profiles and within-profile covariances fixed at zero. Model selection jointly considered fit, profile size, classification quality, substantive distinctiveness, and covariance-specification sensitivity. Associations with profile membership were examined using design-adjusted, classification-error-corrected three-step analyses.ResultsA five-profile solution was retained as a parsimonious descriptive classification: Low Overall / Very Low Confidence, Low Liking-Value / Relatively Higher Confidence, Slightly Below-Average, Positive, and Highly Positive. Perceived school belonging was the most consistent correlate of profile membership, F(4, 68) = 47.74, p < 0.001, and home educational resources also differentiated profiles, F(4, 68) = 15.61, p < 0.001. Gender showed a smaller omnibus association, F(4, 68) = 2.88, p = 0.029, whereas bullying victimization did not differentiate profiles after simultaneous adjustment for the other covariates, F(4, 68) = 1.78, p = 0.143. When perceived school belonging was omitted in a sensitivity model, bullying victimization did differentiate profiles, F(4, 68) = 6.02, p < 0.001.DiscussionMathematics attitudes varied mainly in overall level, with limited configurational heterogeneity among the two lowest-attitude profiles. The change in the conditional association with bullying after omitting school belonging does not establish mediation or causal ordering.