Combination of Unreliable Evidence in a Probabilistic Visual Search and Detection Task.

摘要

Observers watched a display screen to detect a target occurring in 0.1 of the display intervals, with or without the assistance of a simulated unreliable computer. The computer's errors were independent of the random variability in the display, so that assisted subjects had in principle more information upon which to base judgments. Observers with assistance detected more targets than those without it, but also made more false alarm reports so that the benefit overall was slight.