Results chains should be composed of assumptions that can be tested.

Clear­ly state assump­tions so that the strength of evi­dence for each of the assump­tions and link­ages between assumed inter­me­di­ate out­comes in your results chains can be assessed. For exam­ple, ask “What has to be true in order for the pro­posed rela­tion­ship between node A and node B to exist?” In the wind ener­gy exam­ple, the first link assumes that if util­i­ties require avoid­ance in sen­si­tive areas, then they will only buy pow­er from com­pa­nies that have avoid­ed sen­si­tive areas.