Machine Learning Analysis of Emotional Appeals in Charity Crowdfunding

Charity crowdfunding campaigns rely on emotional appeals to elicit empathy-driven support, yet they often struggle to translate such appeals into fundraising success. Drawing on the elaboration likelihood model, this study examines how emotions conveyed through textual descriptions and facial images relate to fundraising outcomes in charity crowdfunding. Using data from a leading crowdfunding platform, we apply machine learning techniques to analyze emotional content across both textual and visual modalities. Our analyses indicate that emotions expressed in textual descriptions contribute more to predicting fundraising performance than emotions conveyed through facial images, consistent with the higher levels of cognitive elaboration typically associated with text processing. Notably, our results reveal that anticipation in text is negatively associated with the amount raised, whereas sadness conveyed through facial expressions is positively associated with fundraising outcomes. To interpret these patterns, we introduce affective modality alignment, emphasizing that emotional appeals tend to be more effective when aligned with the processing tendencies of their delivery modality. By integrating machine learning with the ELM framework, this study advances understanding of emotion-based persuasion in digital philanthropy. Practically, it provides guidance for designing charity crowdfunding campaigns in cultural settings influenced by collectivist norms and offers insights that extend to similar institutional and sociocultural contexts.