This study investigated sex differences in adult autism screening by comparing self-reports on current traits (Autism Spectrum Quotient
AQ) and parent-reports on childhood traits (Relatives Questionnaire
RQ). The aim was to examine the differential contribution of these distinct measures to diagnostic classification in both sexes. The study compared 102 clinically diagnosed autistic adults (30 females) and 152 non-autistic adults (60 females), aged 17-35 years. Participants completed the AQ, while their parents completed the RQ. Multivariate analysis of variance and sex-stratified discriminant analyses were employed to evaluate measurement patterns in males and in females. Significant main effects were found for diagnostic group on both measures and for sex on the RQ only. Group × sex interactions were significant for both measures. Within the autistic group, males and females showed no significant difference in AQ scores, but females scored significantly lower than males on the RQ. Discriminant analyses revealed high classification accuracies for both males (95.1%) and females (96.7%), with different weighting patterns between males (AQ = 0.597 [CI: 0.413-0.720], RQ = 0.712 [CI: 0.553-0.789]) and females (AQ = 0.763 [CI: 0.637-0.898], RQ = 0.478 [CI: 0.191-0.616]). The findings suggest that current self-report may be more central for identifying autism in females, while a more balanced combination of current-self and past-parent reports may be optimal for males. These sex-specific patterns highlight the importance of considering both current self-reported traits and developmental history in adult autism screening, with potential implications for improving diagnostic accuracy across sexes.