Skip to Main Content
 

Global Search Box

 
 
 
 

Files

File List


Dissertation_final.pdf (14.79 MB)

ETD Abstract Container

Abstract Header

Dynamics of Multi-attribute Decision Making Revealed by Eye-tracking

Liu, Qingfang
http://orcid.org/0000-0003-2425-7655
http://rave.ohiolink.edu/etdc/view?acc_num=osu1609933430042674

Abstract Details

2021, Doctor of Philosophy, Ohio State University, Psychology.
Eye-tracking has been a useful process-tracing measure to disclose attentional mechanisms in human decision making, such as in multi-attribute decision making. In recent years, the field of multi-attribute decision making has been greatly advanced by the development of formal computational models. However, most computational models of multi-attribute decision making only deal with choices and response times. To date, there has been no detailed investigation on combining eye-tracking measures with computational models in multi-attribute decision making. This dissertation aims to investigate possible approaches that integrate eye fixation data with computational models, and to provide new insights into decision dynamics underlying multi-attribute decision tasks. Specifically, I adopted and evaluated two different approaches to combine computational models with eye fixation data in intertemporal choice and simple risky choice. The first fixation-modulated approach feeds eye fixation data into sequential sampling models as exogenous inputs, and predicts behavioral responses based on combinations of feature values and eye fixations. The second generative modeling approach predicts both choices and eye fixation data simultaneously through a single model structure with three essential cognitive components - information sampling, feature representation, and preference formation. Different model configurations under the generative modeling approach were constructed for both intertemporal choice and simple risky choice, where different models hold different assumptions about feature representation in multi-attribute tasks. These models were quantitatively fitted to experimental data from Amasino et al. (2019) and Stewart et al. (2016), with the synthetic likelihood approximation method and a set of summary statistics, in a Bayesian framework. The models displayed good fits to the experimental summary statistics. The model comparison results suggested large individual variability regarding the relative performance of different model configurations, where the attribute-wise generative model showed a slight advantage when aggregating across subjects, for both intertemporal choice and simple risky choice. Although there is still room for improvement and generalization, the current work has opened a novel way to successfully model behavioral and eye fixation data in multi-attribute decision making.
Brandon Turner (Advisor)
Alexander Petrov (Committee Member)
Ian Krajbich (Committee Member)
189 p.
Psychology
multi-attribute decision making; attention; eye-tracking; computational model; Bayesian model

Recommended Citations

Citations

  • Liu, Q. (2021). Dynamics of Multi-attribute Decision Making Revealed by Eye-tracking [Doctoral dissertation, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1609933430042674

    APA Style (7th edition)

  • Liu, Qingfang. Dynamics of Multi-attribute Decision Making Revealed by Eye-tracking. 2021. Ohio State University, Doctoral dissertation. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1609933430042674.

    MLA Style (8th edition)

  • Liu, Qingfang. "Dynamics of Multi-attribute Decision Making Revealed by Eye-tracking." Doctoral dissertation, Ohio State University, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=osu1609933430042674

    Chicago Manual of Style (17th edition)

osu1609933430042674
304
© 2021, all rights reserved.
This open access ETD is published by The Ohio State University and OhioLINK.