Title: Perceived Credibility in Visualizations

Date: Monday, May 13th

Time: 1:00 pm - 3:00 pm EST 

Location: Tech Square 334

Zoom: https://gatech.zoom.us/j/94590451441

Hayeong Song

PhD Candidate in Computer Science 

School of Interactive Computing                    

Georgia Institute of Technology

Committee: 

Dr. John Stasko (advisor) – School of Interactive Computing, Georgia Institute of Technology

Dr. Alex Endert – School of Interactive Computing, Georgia Institute of Technology

Dr. Cindy Xiaong – School of Interactive Computing, Georgia Institute of Technology

Dr. Clio Andris – School of Interactive Computing, Georgia Institute of Technology

Dr. Danielle Szafir – Department of Computer Science, University of North Carolina

Abstract:

Well-designed visualizations can leverage the strength of our perceptual capabilities and augment viewers’ cognition to find insights about data, facilitate content comprehension, and enable informed decision-making. However, poor visualizations can obstruct understanding of the content and can even bias viewers’ data interpretation and analysis. In this context, visualization serves as a medium between viewers and the information being conveyed, enhancing credibility in both the underlying data and the communicated message. This dissertation focuses on developing a better understanding of how combinations of different design choices affect viewers' perceived credibility in visualization.

In the context of visualization, we can evaluate credibility in two aspects, 1) the data being represented and 2) the message of a visualization. Regarding the data itself, while people may expect a dataset to be complete and error-free, sometimes data can be incomplete (e.g., uncertainty, missing attributes, or values). For a visualization to be credible, it should be transparent about the quality of underlying data. If the quality is not communicated, it can bias viewers' data interpretation and degrade their perceptions of the data credibility. This dissertation addresses this gap by investigating and studying the effects of visualization factors – missing data imputation methods and visualization techniques- that shape viewers' perceptions of credibility.

In the context of evaluating message credibility, assuming that the data is complete and error-free, how does visualization design affect message credibility? For example, visual embellishments could make a chart memorable and engaging, but they also could undermine the credibility of a message by diminishing the seriousness of the conveyed message. Although we speculate that visual embellishments may potentially erode message credibility, there is no empirical evidence to support that claim. To understand this effect, this dissertation studies how differently styled (embellishments) visualizations affect people's perceived message credibility and which visualization characteristics (chart elements) make an impact. Our findings suggest ways that visualizations might leverage embellishment to effectively communicate engaging messages without degrading perceived message credibility.