to facilitate and support efforts to fulfill tasks.

A series of controlled evaluations across three Fifth Dimension programs, including pre-post assessments and a controlled quasi-experimental design, showed significant gains across a range of outcomes. Studying students after 10-20 sessions, Mayer and colleagues (1997) found positive outcomes for computer skills (knowledge of new terms and facts, operating a computer), reading comprehension, and problem-solving skills. Although there were no science-specific learning outcomes, the broad range of positive results suggests that the program design is promising.

One other feature of the program deserves mention here. The Fifth Dimension Program succeeded in creating an after-school environment in which heterogeneous groups regularly engage in joint, dialogic problem solving. Building these cognitively rich activities into a program organized on a drop-in basis is extremely rare and can be quite difficult to create, even in classrooms in which training levels and other resources are more abundant. Brown and Cole (1997) attribute this success to the social structure—children, undergraduates as helpers and co-participants, and the computer-based Wizard—which decentralizes authority and invites and supports participants’ curiosity and sustained engagement.

Future research and development could examine which elements of this approach could be emulated in science-specific programming and specifically test ways of structuring learner and facilitator roles to build productive, engaged scientific inquiry.