In this conceptual paper, we propose an heuristic to balance context-specific and generic scaffolding, as well as computer-based and teacher scaffolding, during instruction centered on authentic, scientific problems. This paper is novel in that many researchers ask a dichotomous question of whether generic or context-specific scaffolding is best, and fail to focus on what processes and cognitions each type of scaffolding excels at supporting. To arrive at the heuristic and construct the framework, we synthesized research on (a) student challenges gaining and using scientific reasoning strategies, and (b) computer-based and teacher scaffolding.