Public participation theory assumes that empowering communities leads to enduring support for new initiatives. The New Zealand Biodiversity Strategy, approved in 2000, embraces this assumption and includes goals for community involvement in resolving threats to native flora and fauna. Over the last 20 years, community-based ecological restoration groups have proliferated, with between 600 and 4000 identified. Many of these groups control invasive mammals, and often include protection of native species and species reintroductions as goals. Such activities involve the groups in "wicked" problems with uncertain biological and social outcomes, plus technical challenges for implementing and measuring results. The solution might be to develop a citizen science approach, although this requires institutional support. We conducted a web-based audit of 50 community groups participating in ecological restoration projects in northern New Zealand. We found great variation in the quality of information provided by the groups, with none identifying strategic milestones and progress towards them. We concluded that, at best, many group members are accidental scientists rather than citizen scientists. Furthermore, the way community efforts are reflected in biodiversity responses is often unclear. The situation may be improved with a new approach to data gathering, training, and analyses.