Name: Sweta Parmar 

Dissertation Proposal Meeting 

Date: Monday, June 27th, 2022 

Time: 2 PM 

Location: https://gatech.zoom.us/j/93844361599?pwd=MjdFMW5zMWNHMnhDQ1AzdjBjUmNxQT09 or JS Coon 148 

Advisor: Rick Thomas, Ph.D. (Georgia Tech) 

Dissertation Committee Members: 

Jamie Gorman, Ph.D. (Georgia Tech) 

Sashank Varma, Ph.D. (Georgia Tech) 

Karen Feigh, Ph.D. (Georgia Tech) 

Elizabeth Whitaker, Ph.D. (GTRI) 

Title: Model Blindness: Investigating a Model-Based Route-Recommender System’s Impact on Decision Quality  

Abstract: Model-Based Decision Support Systems (MDSS) are prominent in many professional domains of high consequence, such as aeronautics, emergency management, military command and control, healthcare, nuclear operations, intelligence analysis, and maritime operations. An MDSS generally uses a simplified model of the task and the operator to impose structure to the decision-making situation and provide information cues to the operator that is useful for the decision-making task at hand. Models are simplifications, can be misspecified, and have errors. Adoption and use of these errorful models can lead to the impoverished decision-making of users. I term this impoverished state of the decision-maker model blindness. This dissertation aims to investigate the detrimental consequences of model blindness on human decision-making and performance and how those consequences can be mitigated through a series of two experiments. The proposal also reports simulation results that motivate these experiments by demonstrating the impact of model blindness and model blindness mitigation strategies on performance. The proposed experiments will implement a simulated route recommender system as an MDSS with a true data-generating model (unobservable world model). In Experiment 1, the true model generating the recommended routes and the associated attribute information will be misspecified to different levels to impose model blindness on MDSS users. In Experiment 2, the same route-recommender system will be employed with a mitigation technique to overcome the impact of model-misspecifications on decision quality. This dissertation will provide useful insights on how model blindness caused by model misspecification can manifest in performance and decision quality and how to mitigate this state of the decision-maker. This will also help establish a need for evaluating model blindness during the development, implementation, and usage stages of MDSS.