Abstract
People remembering the spatial location of briefly displayed objects use categorical information about scenes to reduce errors in judgment (Huttenlocher, Hedges, & Duncan, 1991). People are thought to adjust inexact memory for the object's location with information about the region (e.g., "right half of the screen") in which it appeared, leading estimates to be biased toward category prototypes. The categorical organization used is thought to be consistent across people and various task manipulations. In contrast, the current study suggests that category structure may change as a function of memory delay. Undergraduates remembered the location of single dots in a large rectangular display for either 300ms or 3000ms. Then a centrally located mouse pointer appeared, and participants moved it to indicate their remembered location. Observed bias was large at the longer delay. Responses tended to be biased outward from the center across much of the range and inward near the outside of the screen. In contrast to previous work (Huttenlocher et al., 1991), estimates for short-delay trials were biased outward across most of the range, decreasing only due to truncation at the screen edge. Though responses at short delays showed a clear systematic deviation from the stimulus location, the form of the bias is not compatible with a category adjustment model that assumes a category prototype on the screen. At short delays, the bias appears to be relational: dots were categorized as left or right, but not drawn to a central, prototypic value. In contrast to prior findings, these results indicate that manipulations of temporal delay (and thus memory uncertainty) lead people to take advantage of different types of categorical information.
Meeting abstract presented at VSS 2014