Background: Sexual minority individuals consistently report higher rates of mental disorder than heterosexuals. However, much of the research has methodological limitations related to the classification of sexuality, the use of cross-sectional data and problematic sampling procedures such as using convenience samples. Methods: We used longitudinal data from a birth cohort enrolled in the Christchurch Health and Development Study (n = 1040). Latent class analysis was used to classify participants sexuality based on self-report data of sexual behaviour, attraction, identity and fantasy, gathered over five assessments between the ages of 18 and 35 years. Mental health and substance use outcome data were gathered at four assessments between the ages of 21 and 35 years. Potential covariate variables were collected during childhood. Results: The latent class analysis identified four groups interpreted as: 'heterosexual' 82%, 'mostly heterosexual' 12.6%, 'bisexual' 3.5% and 'gay/lesbian' 1.9%. In the sexual minority groups, women outnumbered men by at least 2:1. Pooled rates for mental health disorders of depression, anxiety disorders, suicidal ideation, cannabis abuse and total disorders, after adjustment for childhood covariate variables, were significantly higher in the sexual minority groups (p < 0.01). The strength of association between sexuality group and mental health outcomes did not differ according to sex. Fluidity in sexuality reports appeared unrelated to risk of mental health outcomes. Conclusions: Over the life course, membership of a sexual minority group is clearly associated with mental health problems of depression, anxiety and suicidal ideation regardless of the age when same-sex attraction, behaviour, identity or fantasy is expressed. [ABSTRACT FROM AUTHOR]