Just say 'I don't know': Understanding information stagnation during a highly ambiguous visual search task.

Godwin, Hayward J; Hout, Michael C · 2023 · DOAJ

DOI: 10.1371/journal.pone.0295669

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study investigates visual search behavior and decision-making when participants are provided with a third response option, “I don’t know,” in addition to the standard “present” or “absent” choices. Traditional visual search experiments typically use simple, easily identifiable stimuli, rendering an uncertainty option unnecessary. However, real-world search tasks, such as radiology or airport security screening, often involve ambiguous displays where observers may be unable to definitively categorize a target. The authors aim to fill a theoretical gap by examining how searchers handle ambiguity, specifically testing the concept of “information stagnation,” where insufficient visual information prevents a decision between target presence or absence. The researchers conducted an online experiment with 277 undergraduate participants, who searched for the letter ‘o’ amidst overlapping distractor letters. The study employed a between-subjects design with two conditions: a Standard Search condition (two response options) and an “I Don’t Know” Search condition (three response options). Display difficulty was manipulated by varying the degree of overlap among letters into low, medium, and high levels. Participants completed 270 experimental trials. Data were rigorously cleaned to exclude disengaged participants, resulting in a final dataset of 45,951 trials from 173 participants. The authors analyzed response rates and reaction times (RTs) using Generalized Linear Mixed Models (GLMMs). The results confirmed that as overlap increased, participants in the “I Don’t Know” condition were significantly more likely to select the uncertainty option. Reaction time analyses revealed distinct patterns based on overlap levels. In low-overlap displays, “I don’t know” responses occurred later than “present” responses but earlier than “absent” responses. Conversely, in high-overlap displays, “I don’t know” responses were generated very rapidly, often faster than “present” responses. This suggests that when displays are highly ambiguous, searchers quickly encounter clusters that trigger information stagnation, leading to immediate uncertainty responses rather than prolonged search efforts. Additionally, response accuracy was higher in the Standard Search condition, likely because participants were forced to guess rather than opting out. These findings challenge existing models of visual search, such as the Guided Search model, which do not account for uncertainty responses. The study demonstrates that “I don’t know” decisions are not merely late-stage failures but can be rapid responses triggered by local ambiguity. This has significant implications for understanding human performance in high-stakes, ambiguous search environments, suggesting that providing an uncertainty option allows observers to efficiently flag difficult stimuli rather than expending resources on futile searches. The concept of information stagnation provides a new framework for modeling how visual search systems handle insufficient evidence.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-25; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• visual search