Misinformation on social media is a phenomenon with considerable impacts on our societies. To effectively mitigate its effects, we need to employ a cross-disciplinary effort to holistically address the critical elements involved in the process, including the source, the content, and the consumers of misinformation. However, the vast majority of computational approaches to addressing this problem focus on detecting and flagging misinformation content. Other important aspects, such as behaviors and intents of sources and the consumers’ decision-making processes, are often neglected. In this thesis, we address this gap through a series of studies around how users make decisions about content and sources of misinformation facilitated by Visual Analytics. First, We introduce a Visual Analytic system, Verifi, that combines temporal, language, and network analysis features and enables users to assess the veracity of multiple news sources. Using Verifi, we conducted a controlled experiment that highlighted how uncertainty and conflicting cues in information impacts users’ perceptions of source credibility. Next, we extended Verifi to a more comprehensive multi-modal system enabling users to study sources through social network analysis, text, and images. Through a qualitative domain expert study conducted on Verifi, we learned valuable lessons about the importance of users’ trust in sources and how emotional content in images might impact users’ judgments and perceptions. Inspired by the lessons learned from our work on Verifi, we present two studies on the effects of emotions in images on users’ perception of content bias and source credibility. In the first study, we investigated the impact of happy and angry portrayals in facial imagery on users’ decisions. In the second study, we explore the interaction between users’ prior attitude towards multiple personalities and how those personalities are portrayed visually on their decisions. Our results show that the systematic usage of angry facial emotions in images increases users’ perceived content bias and decreases the perceived source credibility. These results highlight how implicit visual propositions by news sources impact our judgements and pave the way for visual analytic systems that are sensitive to users' individual and group interactions with such visual information.