Craig P. Paukert

River restoration programmes with the goal of conserving and rehabilitating inland fishes have a multi‐decadal history, but evaluation and synthesis of past restoration actions have been limited by a lack of monitoring and reporting. Given that calls for both monitoring and systematic reviews of restoration have increased, we were interested in the influence that restoration has had on improving conditions for riverine fishes resulting from long‐standing and increasingly prominent stressors. Our objectives were to (1) identify which stressors were targeted in river restoration efforts, (2) determine the proportion of published studies that articulate restoration goals and develop comparative monitoring designs for assessing the effects of restoration on fishes, and (3) conduct a meta‐analysis to synthesise fish responses to restoration projects. We assessed restoration effectiveness for increasing fish density and richness from peer‐reviewed papers published over the past decade using a global systematic review and meta‐analysis. We found that restoration actions addressed major stressors primarily by improving in‐stream habitat (37%), increasing in‐stream longitudinal connectivity (26%) and increasing lateral floodplain connections (9%). Although 81% of studies had comparative monitoring designs (i.e., before/after and control/impact) and stated restoration goals, only 40% of those studies reported sufficient data to be included in the meta‐analysis. Projects which increased in‐stream connectivity had the largest positive effect size on fish density and richness compared to in‐stream habitat improvements and increasing floodplain connections. Time since restoration and restoration size (i.e., geographical footprint) were not strong predictors of fish response effect sizes. Restoration effectiveness was highly variable among project types. Authors of studies included in the meta‐analysis often identified spatial or temporal scale of monitoring, overriding catchment conditions, and recolonisation potential as sources of variability and effectiveness in restoration outcomes. Systematic reporting of these and other covariates may help guide processes in restoration evaluation and provide valuable research insights. Despite increased emphasis on monitoring, incomplete data reporting limited the number of studies that could be included for quantitative meta‐analysis. Persistent emphasis on setting specific criteria (e.g., measurable outcomes of fish response) for restoration goals, project monitoring, data reporting, information sharing and collaborative projects is likely to continue to improve understanding of restoration effectiveness transferable to future endeavours. Our results can be used to support river restoration practitioners with evidence‐based information to evaluate the cost–benefit ratio of competing restoration priorities, and inform restoration planning and implementation for riverine fish.