The nature of comparison conditions is a much overlooked feature of designing and interpreting the results of randomized controlled trials, as outlined by Williams (see record 2010-18776-001). We agree that understanding the components and mechanisms of the comparison condition is necessary for making inferences about both intervention effectiveness (whether the intervention worked and which components may have contributed to such an effect) and about theoretical mediators (how it worked). The extent to which one can draw strong inferences regarding the efficacy and mechanisms of an intervention over the comparison is conditional upon a number of key points. (PsycINFO Database Record (c) 2010 APA, all rights reserved)