This chapter explores reasons and mechanisms behind the desituating of data in aircraft accident analysis, including retrospection, the social construction of cause, and investigators' susceptibility to lay interpretations. In particular, two patterns in desituating data are identified, namely micromatching and cherry-picking. By micro-matching is meant linking behavioral fragments with after-the-fact-worlds. In cherry-picking, the investigator gathers fragments from a wide variety of places and times in the accident record--not because these fragments necessarily belong together but because they all support an argument that the investigator wants to make. The fundamental surprise error means that large, complex events get picked apart and reinscribed to fit existing assumptions about the system in which they took place--rather than that our assumptions get changed so that they can accommodate the possibility of failure. Aircraft accident analysis contributes to this error through retrospective desituating of data, proximal focus, and counterfactual and judgmental language. As a result, different and mutually exclusive investigations can emerge, each emphasizing and relying on its own sets of cherries as derived from a single sequence of events. The result unfortunately leaves the larger airline system or industry none the wiser.