Not in the Picture: Visual Descriptive Representation and Gender Gaps in Parties’ Visual Communication

Political communication has increasingly shifted toward visual and social media platforms, reshaping how parties engage with voters and, crucially, what they show them. As political competition becomes ever more image-driven, parties now curate visual content to project inclusivity, modernity, and responsiveness. Yet, while research on descriptive representation has focused on institutional inclusion, how many women hold office or leadership positions, little attention has been paid to whether these groups are actually visible in party-controlled imagery. This paper introduces the concept of visual descriptive representation, defined as the extent to which social groups are visibly represented in the visual communication that parties fully control. Using gender as a central case, we examine how political parties visually present women across their Instagram communication. Our analysis draws on an original dataset of over 600,000 images posted by 360 political parties in 38 countries, covering each account from its creation through 2023. Employing computer vision techniques, we automatically detect faces and classify gender, allowing us to measure women’s visual presence at scale and compare it to their institutional descriptive representation in parliaments. The results reveal a consistent and substantial gender gap: women are systematically underrepresented in parties’ visual communication, even within parties with high institutional gender representation or female leadership. Across countries and party families, no major party reaches gender parity. Left-wing parties feature women more prominently, while conservative and radical-right parties display the lowest levels of female visibility. Moreover, visual representation often diverges strategically from institutional patterns. Some parties visually overrepresent women, signaling inclusivity at low cost without structural change, while others underrepresent them despite substantial institutional inclusion, revealing persistent masculine norms of political visibility. These findings expand the concept of descriptive representation beyond formal officeholding to the realm of public visibility. They demonstrate that even in the low-cost, flexible sphere of social media, gender inequality endures in the symbolic presentation of politics. Representation, therefore, is not only about who is elected but also about who is seen. Visual descriptive representation exposes a new, image-based dimension of democratic inequality in the digital age.