Social media platforms' ubiquity has fundamentally reshaped how individuals communicate, share information, and engage with one another in the digital age. This dissertation investigates the complex dynamics of online information sharing, focusing on the role of emotions in these processes. Through three interconnected studies, this work contributes to understanding how emotions shape online interactions and influence the spread of information.The first study extends Pennebaker's (1993) offline stage model in the aftermath of large-scale upheavals to the online environment, exploring the temporal dynamics of emotional expressions on social media in the aftermath of earthquakes. By analyzing Twitter data from three events and using methods such as change point analysis, analysis of variance, time series clustering, and Granger-causality tests, the study identifies different stages of online communication that align with Pennebaker's model. The findings highlight the importance of considering the temporal dimension in understanding online emotional dynamics and provide insights into the role of low arousal emotions as drivers of information sharing in crisis communication. The second study investigates the role of emotions in the information diffusion of social media content, focusing on the differences between lean-back (passive) and lean-forward (proactive) sharing behaviors. Through the analysis of six case studies across political affairs, natural disasters, and sporting events on Twitter, the study combines Natural Language Processing techniques and Structural Causal Modeling to understand the differences between lean behaviors. The findings reveal that the arousal dimension of emotions is more relevant than valence in distinguishing between lean behaviors and emphasize the influence of the topics on the emotional differences between these behaviors. Moreover, the study detects the emergence of emotional synchronization within information cascades, highlighting the importance of considering lean behaviors simultaneously when studying online information sharing. The third study examines the relationship between context (mood), content (credibility), and individuals' emotions to understand the online misinformation sharing of short videos. Utilizing a preregistered survey experiment with a single-factor design, the study demonstrates that mood and emotions play a joint role in the sharing process, with positive mood having both direct and indirect effects on sharing behavior, while negative mood influences sharing behavior mediated by negative emotions. The findings also suggest that more credible misinformation has a stronger impact on the relationship between negative emotions, moods, and content sharing than less credible information. While existing literature has established a relationship between emotions and information sharing, particularly in describing the types of emotions that make content viral (typically high-arousal negative emotions), there is a gap in understanding the complementary role of emotions as both causes and effects of information sharing. This dissertation contributes to filling this gap by conceptualizing the information sharing behavior as a flow generated by emotions, which also produces content that evokes emotions. The results of this research uncover the significant role of low arousal emotions, such as sadness, interest, and disgust, as drivers in information sharing. These findings show that these emotions can trigger the expression of higher-arousal emotions like anger, revealing a nuanced mechanism in which low arousal emotions play a latent role, motivating content sharing actions that complement the effects of high arousal emotions. Understanding these emotional interactions is crucial for explaining the complex pathways through which different emotional states contribute to spreading information within the digital ecosystem. This dissertation uses a comprehensive methodological approach, combining computational methods, observational data, and experimental design with causal inference analyses to provide a thorough understanding of the role of emotions in online communication and information diffusion.