Abstract
In a sustained inattentional blindness task, people who selectively attend to one set of objects and ignore another fail to notice unexpected objects. Their likelihood of noticing is driven by similarity of the unexpected object to the attended and ignored sets, but how do we assign membership to these sets? Do we select individual object features (e.g., enhancing "black" and suppressing "white") or do we group objects into categories like "white" and "non-white?" These two accounts predict different rates of noticing for unexpected objects that share a set's category but not its member features. In a large, preregistered study (https://osf.io/7pz35), participants tracked either a set of four white shapes or a set containing four uniquely colored shapes (red, yellow, black, and purple), counting the number of times the attended items bounced off the edges of a displayed window. On the third trial, an unexpected shape passed though the display. The shape was either white, a color in the nonwhite set, or a color unique to the display but in the "nonwhite" category (green). If attention sets are category-based, green objects should be ignored along with the red, black, purple, and yellow objects because they too fall into the "non-white" category. Consistent with this prediction, when ignoring the nonwhite shapes, participants noticed the green unexpected object no more often than one exactly matching an ignored color, and when attending to the nonwhite shapes, participants were just as likely to notice a green shape as one with an attended color. In this task, attention sets were based on broader object categories rather than individual object features. Noticing of an unexpected object depends on which category it belongs to, and not merely how similar its features are to the attended and ignored sets.
Meeting abstract presented at VSS 2017