This was not a prospective study, but instead was conducted by survey - the least reliable measurement instrument that I am aware of. Nevertheless, the authors

This was not a prospective study, but instead was conducted by survey - the least reliable measurement instrument that I am aware of. Nevertheless, the authors didn't do too much begging of the question. They used the term 'linked' judiciously, and admitted that the disease process already seemed to have started before commencement of marijuana use. Had they claimed that marijuana use causes psychosis, I would be screaming "Ascertainment bias!" But they refrained; so shall I.