Do You Trust What You See? Toward A Multidimensional Measure of Trust in Visualization

TL;DR

This paper explores the complex concept of trust in information visualization, identifying key factors influencing trust and empirically examining how design features affect user perceptions and trustworthiness.

Contribution

It introduces a multidimensional measure of trust in visualization, highlighting five key factors and empirically validating their impact on user trust perceptions.

Findings

Credibility, clarity, reliability, familiarity, and confidence influence trust.

Visualization design and source affect trust ratings.

Trust factors align with subjective trust rankings.

Abstract

Few concepts are as ubiquitous in computational fields as trust. However, in the case of information visualization, there are several unique and complex challenges, chief among them: defining and measuring trust. In this paper, we investigate the factors that influence trust in visualizations. We draw on the literature to identify five factors likely to affect trust: credibility, clarity, reliability, familiarity, and confidence. We then conduct two studies investigating these factors' relationship with visualization design features. In the first study, participants' credibility, understanding, and reliability ratings depended on the visualization design and its source. In the second study, we find these factors also align with subjective trust rankings. Our findings suggest that these five factors are important considerations for the design of trustworthy visualizations.