Memory Theoretic Account of Human Information Foraging
Author(s)
Youn, Anne
Advisor(s)
Editor(s)
Collections
Supplementary to:
Permanent Link
Abstract
Human decision making often requires people to consider several possible explanations, revise those explanations as new evidence appears, and decide when enough information has been gathered to act. This is essential in diagnostic reasoning, where decision makers must evaluate competing hypotheses while deciding on which information is worth seeking next. Information foraging theory explains how people search for useful evidence, especially when some sources of information are more diagnostic or informative than others. However, this approach does not fully explain how internally generated hypotheses, memory constraints, and subjective confidence shape the decision to continue searching or stop.
This thesis builds on the idea of hypothesis-guided search: early hypotheses do not simply reflect the decision maker’s beliefs, but also shape which information sources are considered valuable and when search appears worth terminating. Through Python code, this thesis aims to develop a psychologically plausible representation of human information foraging. The developed computational model will test the impact of the composition of hypotheses, time pressure, working memory capacity, and cost sensitivity on information search and search termination behavior.
More importantly, it will explore the hypothesis that human hypothesis generation and other information processing constraints specified by the theory shape information search and search termination behavior. By combining the trends observed through the various studies on hypothesis generation, this thesis hopes to create a more holistic, thorough, and comprehensive approach to this subject. A better understanding of the topic of hypothesis generation could be the key to understanding why humans create decisions and generate hypotheses the way they do, which is an essential subject in all areas of study- since every field has hypothesis generation.
Sponsor
Date
2026-05
Extent
Resource Type
Text
Resource Subtype
Undergraduate Thesis