Author(s) :

Yufei Xiao.

Volume/Issue :

Abstract :

Autism spectrum disorder (ASD) is a complex neurodevelopmental condition characterized by differences in communication, behavior, and cognition, with significant social and familial implications. Understanding how familial background contributes to autism-related behavioral traits remains an essential research goal. This study utilizes a publicly available dataset from Kaggle, comprising more than 700 respondents, to examine the relationship between family history of autism and ten standardized screening items (A1–A10). Two complementary statistical frameworks were employed: individual binary logistic regression models, which estimate the item-level association between family history and each screening response, and a joint Generalized Estimating Equation (GEE) model, which accounts for within-subject correlation among multiple items. Results from the logistic regressions reveal significant positive associations for six items (A1, A3, A4, A5, A6, A9, and A10), with odds ratios for these significant items ranging from approximately 1.7 to 3.5, clarifying that non-significant items (e.g., A2 = 1.56, A8 = 1.25) are not included in this range. The joint GEE analysis further confirms an overall odds ratio of 1.86 (p < 0.001), indicating that participants with a family history of autism are nearly twice as likely to respond positively to ASD-consistent screening indicators. Together, these findings provide statistical evidence for a familial component in autism-related behavioral expression and demonstrate the value of integrating individual and joint modeling techniques in autism data analysis.