Literature on the relationship of cannabis use and cognition in schizophrenia provides the paradoxical view that cannabis use is sometimes linked with less severe impairment in neurocognition. This paper explored the possibility that this is a reflection of a dose related response between lifetime cannabis use and two forms of cognition, neurocognition and metacognition, in schizophrenia. It was hypothesized that three groups of patients could be differentiated, those with (1) little to no cannabis use with poor levels of cognition, (2) moderate cannabis use and relatively better levels of cognition and (3) high cannabis use with relatively poorer levels of cognition. Sixty-six adults with schizophrenia completed assessments of neurocognition, metacognition and months of lifetime cannabis use. A k-means cluster analysis yielded three distinct groups based on these assessments. The clusters included: (1) low cannabis/poor cognition (n = 34); (2) heavy cannabis/moderately impaired cognition (n = 10); and (3) moderate cannabis/higher cognition (n = 22). Consistent with our hypothesis, participants with high and moderate lifetime cannabis use had lesser impairment of neurocognition and metacognition compared to low lifetime cannabis use. Participants with moderate lifetime cannabis use also had lesser impairment of metacognition compared to low and heavy use. These findings suggest that a dose related relationship exists between cannabis use and cognition. Results could be due to an influence of pre-existing cognitive level on likelihood of lifetime cannabis use, or to an interaction between use and cognitive function.