Sexuality- and gender-diverse youth experience disproportionate health and social adversity. Accurate early-life indicators are important for development of supportive approaches. To examine whether commonly used items measuring childhood conformity to gender roles are associated with sexual orientation in young adulthood. This single-center, prospective cohort study (the Raine Study) assessed 2868 children of 2900 women who were recruited during pregnancy from August 1, 1989, to April 30, 1992, with follow-up ongoing. The Achenbach System of Empirically Based Assessment (ASEBA) Child Behavior Checklist, Teacher Report Form, and Youth Self-Report tools were used to survey parents, teachers, and youths to examine gender diversity among the participating youths. Parents were followed up at years 5, 8, 10, 14, and 17, adolescents at years 14, 17, and 27, and teachers at years 10 and 14. Data were analyzed from August 1, 2020, to July 31, 2023. Parent and teacher report that a child "behaves like the opposite sex" (gender role behavior), and parent and self-report that a child "wishes to be the opposite sex" (gender role wish), in response to assessment items. Year 27 self-reported sexual identity, attraction, and behavior. Of the 2868 children in the original birth cohort, 1154 (40.2%) participated in the year 27 follow-up, of whom 608 (52.7%) were recorded female at birth and 546 (47.3%) were recorded male at birth. Of these, 582 who were recorded female at birth continued to identify as female (cisgender) (95.7%), and 515 recorded male at birth continued to identify as male (cisgender) (94.3%); 47 (4.1%) did not complete the questionnaire. Of cisgender participants, 76 of 605 women (12.6%) and 52 of 540 men (9.6%) had a diverse sexual identity, 204 of 605 women (33.8%) and 77 of 540 men (14.3%) were same-gender attracted, and 100 of 605 women (18.6%) and 39 of 540 men (7.2%) had ever engaged in same-gender sexual behavior. Across all follow-ups, after adjusting for gender, nonconforming gender role behavior was consistently associated with diverse sexual identity and behavior (adjusted odds ratio [aOR] for identity, 2.8; 95% CI, 1.9-4.2; behavior aOR, 2.4; 95% CI, 1.6-3.5). Self-reported gender role wish was consistently associated with diverse sexual orientation (identity aOR, 2.3; 95% CI, 1.4-3.8; attraction aOR, 1.7; 95% CI, 1.1-2.5; behavior aOR, 1.9; 95% CI, 1.2-2.9). In this cohort study, ASEBA gender role nonconformity was associated with diverse sexual orientation, beginning in early childhood. Findings suggest that the ASEBA measures should not be used to infer sexual orientation or gender diversity in clinical or research settings; asking direct questions may provide more accurate data.