Variability of repeated measurements in human performances exhibits fractal 1/f noise. Yet the relative strength of this fractal pattern varies widely across conditions, tasks, and individuals. Four experiments illustrate how subtle details of the conditions of measurement change the fractal patterns observed across task conditions. The results call into question whether measurement noise and measured signal can be distinguished in human performance, suggesting that human performance is inextricably entangled with measurement context. Perhaps, though, a hypothesis of soft assembly of human performance can circumvent the conundrum (e.g., Turvey, 2007). (Contains 11 figures and 1 table.)