Changes are not localized before they are explicitly detected

Change detection is in many ways analogous to visual search. Yet, unlike search, successful detection depends not on the salience of features within a scene, but on the difference between the original and modified scene. If, as in search, pre-attentive mechanisms guide attention to the change location, the change itself must produce a preattentively detectable signal. Despite recent evidence for implicit representation of change in the absence of conscious detection, few studies have yet explored whether attention is guided to a change location prior to explicit detection. In four ‘‘change blindness’ ’ experiments using several variants of the ‘‘flicker’ ’ task, we tested the hypothesis that implicit or preattentive mechanisms guide change localization prior to explicit detection. None of the experiments revealed improved localization of changes prior to explicit reports of detection, suggesting that implicit detection of change does not contribute to the eventual explicit localization of a change. Instead, change localization is essentially arbi-trary, driven by the salience of features within scenes. When an original and a modified scene are separated by a brief disruption (e.g., a blank screen, an eye movement, etc.), observers are often unable to report the modification, a phenomenon known as ‘‘change blindness’