Prevalence effects on perceptual decisions: Category broadening, elevated miss rates, or both?

Lyu, Wanyi; Levari, David; Nartker, Makaela; Little, Daniel S.; Wolfe, Jeremy M.

doi:10.1167/jov.20.11.720

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While this explanation is speculative, a similar logic was recently argued in a paper examining how low prevalence impacts the decision criterion in a two-alternative classification task (Lyu et al, 2020 ). The authors found that, in the absence of feedback, decision criteria became more liberal, expanding the category boundary of the rare target.…”

Section: Discussionmentioning

confidence: 71%

Quitting thresholds in visual search are impacted by target present detection times but not their variability

Becker

Rodriguez

Pontious

2022

Atten Percept Psychophys

View full text Add to dashboard Cite

Models of visual search posit that target absent responses occur when the quitting threshold for the trial is reached before a target is detected, and that feedback about missed targets allows the quitting threshold to be adaptively set to the difficulty of the search task. While these models may effectively capture processes in lab-based tasks, in real-world searches feedback is often impossible to provide. Instead, observers have little information about their errors, and may only be aware of when they successfully detect the target. We posit that in the absence of feedback the time required to find a target might influence quitting thresholds. In three experiments, we investigate how manipulating the mean time and the standard deviation of time to detect a target influence quitting thresholds in target absent trials. To vary target detection times while holding the search stimuli constant, we used an eye-movement contingent change to surreptitiously introduce a target near fixation at a particular time. Results show that decreasing the mean time to find a target also decreases the number of items inspected and reaction time in target absent trials, the hallmark of a shift in the quitting threshold. By contrast, varying the standard deviation around a fixed mean had no impact on target absent search times. These findings suggest that people are sensitive to the typical time required to find a target in a given task and use that information to flexibly adjust target absent quitting thresholds, but people are not sensitive to the variability.

show abstract

Section: Discussionmentioning

confidence: 71%