“…The MHQ had a very good response rate compared with previous UKB online questionnaires, and it gives an unparalleled sample size for a mental health survey. However, like much observational research, it is subject to participation bias in its volunteers (Davis, Coleman, et al, ; Fry et al, ). Given that participation in research can be patterned by mental health (Atherton, Fuller, Shepherd, Strachan, & Power, ; Knudsen, Hotopf, Skogen, Øverland, & Mykletun, ), it may be that people with severe symptoms of mental disorder were less likely to volunteer or complete the MHQ, as might be suggested by the small number of people with a hospital data‐linkage diagnosis of a psychotic disorder, which may limit generalisability of our findings to other settings.…”