Patient-centred outcomes such as quality of life (QOL) are valued and used extensively in mental healthcare evaluations, but concerns remain about their practical application due to perceived measurement issues, including responsiveness and relationships with objective indicators and depression. Evidence from general population studies challenge some assumptions, suggesting that measurement difficulties might relate to the characteristics of mental health samples, rather than measurement itself. This paper assesses the impact of mental illness on QOL and its measurement, examining whether the life-conditions, opportunities and QOL of different mental health-status groups vary, and if explanatory models of domain-specific and global QOL differ. Objective life-conditions, access to life-opportunities and subjective QOL were assessed over 2 years, using the same methodology in severe mental illness (SMI; n = 149 (baseline)/n = 126 (follow-up)), common mental disorder (CMD; n = 794/354) and no disorder (n = 1119/583) groups. Objective life-conditions were worse in the SMI group than in mentally healthy population and CMD groups, but the opportunities available to the SMI group were no more restricted than the CMD group. Subjective QOL ratings reflected this; SMI group scores were lower than the healthy population and in some life-domains the CMD group. Models of QOL suggested that life-quality was explained differently in the three groups. QOL studies combining mental health samples should control for health-status group, and domain-specific and global indicators of lifestyle and opportunity.