Deficits in social and occupational function are widely reported in psychosis, yet no one measure of function is currently agreed upon as a gold standard in psychosis research. The aim of this study was to carry out a systematic review and meta-analysis of functioning measures to determine what measures were associated with largest effect sizes when measuring between-group differences, changes over time, or response to treatment. Literature searches were conducted based on PsycINFO and PubMed to identify studies for inclusion. Cross-sectional and longitudinal observational and intervention studies of early psychosis (≤5 years since diagnosis) that included social and occupational functioning as an outcome measure were considered. A series of meta-analyses were conducted to determine effect size differences for between-group differences, changes over time, or response to treatment. Subgroup analyses and meta-regression were carried out to account for variability in study and participant characteristics. One hundred and sixteen studies were included, 46 studies provided data (N = 13 261) relevant to our meta-analysis. Smallest effect sizes for changes in function over time and in response to treatment were observed for global measures, while more specific measures of social and occupational function showed the largest effect sizes. Differences in effect sizes between functioning measures remained significant after variability in study and participant characteristics were accounted for. Findings suggest that more specific measures of social function are better able to detect changes in function over time and in response to treatment.