Testing and validating a coherence-based model for decision-making and search

Scharf, Sophie E.

PDF
scharf_dissertation.pdf - Published
Download (4MB)

URL:	https://madoc.bib.uni-mannheim.de/59164
URN:	urn:nbn:de:bsz:180-madoc-591644
Document Type:	Doctoral dissertation
Year of publication:	2021
Place of publication:	Mannheim
University:	Universität Mannheim
Evaluator:	Erdfelder, Edgar
Date of oral examination:	29 March 2021
Publication language:	English
Institution:	School of Social Sciences > Allgemeine Psychologie (Bröder 2010-)
Subject:	150 Psychology 300 Social sciences, sociology, anthropology
Keywords (English):	information search , coherence , parallel constraint satisfaction
Abstract:	Recently, Jekel et al. (2018) proposed the integrated coherence-based decision and search model (iCodes), which predicts decision-making and information search in multi-cue decision tasks. This model assumes that decision-makers strive for coherence-maximization and that this pursuit is represented by an iterative spread of activation through a network that represents all relevant information of the current decision. The goal of my thesis is to evaluate iCodes as a general theory of decision-making and information search, to test its predictions by employing experimental methods, computational modeling and process tracing, and finally to critically discuss its merits in the field of judgment and decision-making research (JDM) as a whole. A unique contribution of iCodes compared to other theories in JDM is its prediction that people show a tendency to search for information on the option that is currently supported by the available evidence, also referred to as the attraction search effect. In the first manuscript, we could show that the attraction search effect is a robust finding and generalizes to a variety of different tasks. This finding supported a broad range of applicability for iCodes’ predictions for information search. In the second manuscript, we could show that iCodes can account for the effect of a theoretically motivated moderator of the attraction search effect, namely that rating the attractiveness of options before search increases the tendency to search for information on the attractive option. We further validated the assumed, underlying information-search process by showing that a model-inherent parameter can account for the effect of attractiveness ratings. The third manuscript showed that iCodes is not only able to predict behavioral information search but is also able to predict gaze behavior in decision-making. It further highlighted the role of coherence for attention allocation in decision tasks. In sum, my thesis contributes to the theoretical advancement of research in JDM and emphasizes the importance of formalized theories.