The use of citizen science to address global and local environmental challenges will depend on demonstrated evidence that it can lead to meaningful contributions to science, management, and social action. Systematic evaluation of citizen science projects is important yet lacking to date. We developed an evaluation tool and used it to conduct a meta-analysis of 51 Earthwatch projects over a 7-year period, assessing their ability to produce peer-reviewed publications and contribute to management plans and policies. The development and testing of an evaluation tool identified key factors to improve outcomes of citizen science projects, including deliberate design of projects through direct engagement with scientists. In turn, scientists increased their reporting of outcomes when outcomes were being used for program assessment and feedback to participants. Over this period, outcomes for the 51 projects consisted of: 333 peer-reviewed publications and 264 contributions to management plans and policies, with a mean of 1.6 peer-reviewed publications per project per year and 1.3 contributions to management plans per project per year. Across this period, projects averaged 6.5 publications and 5.2 contributions to plans and policies per project (range 0–26 contributions per project). Several other project attributes were found to lead to higher outcomes. We found that the creation of evaluation tools helped hold projects accountable for outcomes and highlighted to project managers and scientists the characteristics of projects that lead to improved outcomes. Elements of this approach could be transferred to other projects, helping to fulfill the potential of citizen science to address global challenges.