Date of Award

Spring 5-15-2021

Author's School

Graduate School of Arts and Sciences

Author's Department


Degree Name

Doctor of Philosophy (PhD)

Degree Type



The low prevalence effect (LPE), the tendency for observers to be more likely to miss rare targets than frequent targets, is a robust error and is difficult to reduce. The LPE is an obstacle in a variety of real-world search tasks in which targets are rare, including baggage screening and some medical imaging. The LPE is thought to occur because when an observer searches for a low-prevalence target, over time, the observer may become both more willing to indicate a target is not there and more likely to end the search early. The present experiments employ three selection history effects, including manipulating reward history, action history, and the availability of items in working memory, in order to learn more about the mechanisms behind the LPE, leading to a better understanding of low-prevalence search and possibly reducing the effect. In the reward history task, participants completed a training phase in which a specific feature was rewarded. Then, they completed a search task in which no reward was presented, and the LPE was measured. In the action history task, participants made simple actions or withheld responses as a prime was presented and then completed a search task. In the working memory task, participants held a target in working memory and completed a search task. The availability of low-prevalence targets in working memory reduced the LPE, but action history and reward history did not affect the LPE. The results from these experiments suggest that direct priming of targets is the best way to reduce the LPE, and indirect priming of targets is not as effective. These experiments present the first evidence that selection history affects low-prevalence visual search. Theoretically, this informs both our understanding of visual search at a range of prevalence levels and our understanding of the mechanisms behind the LPE. Practically, these findings may contribute to methods that reduce the LPE in real-world search tasks.


English (en)

Chair and Committee

Richard A. Abrams

Committee Members

Julie M. Bugg, Sandra Hale, Wouter Kool, Zoe Jenkin,