Problem solving is an important skill in the knowledge economy. Research indicates that the development of problem solving skills works better in the context of instructional approaches centered on real-world problems. But students need scaffolding to be successful in such instruction. In this paper I present a conceptual framework for understanding the effects of scaffolding. First, I discuss the ultimate goal of scaffolding--the transfer of responsibility--and one way that scholars have conceptualized promoting this outcome (fading). Next, I describe an alternative way to conceptualize transfer of responsibility through the lens of distributed cognition and discuss how this lens informs how to promote transfer of responsibility. Then I propose guidelines for the creation of problem solving scaffolds to support transfer of responsibility and discuss them in light of the literature. (Contains 1 table and 1 figure.)