GIFs afford a high degree of personalization, as they are often created from popular movie and video clips with diverse and realistic characters, each expressing a nuanced emotional state through a combination of characters' own unique bodily gestures and distinctive visual backgrounds. These properties of high personalization and embodiment provide a unique window for exploring how individuals represent and express themselves on social media through the lens of the GIFs they use. In this study, we explore how Twitter users express their gender and racial identities through characters in GIFs. We conducted a behavioral study (n=398) to simulate a series of tweeting and GIF-picking scenarios. We annotated the gender and race identities of GIF characters, and we found that gender and race identities have significant impacts on users' GIF choices: men chose more gender-matching GIFs than women, and White participants chose more race-matching GIFs than Black participants. We also found that users' prior familiarity with the source of a GIF and perceptions about the composition of the audience (viz., having a matching identity) have significant effects on whether a user will choose race- and gender-matching GIFs. This work has implications for practitioners supporting personalized social identity construction and impression management mechanisms online.
Read full abstract