Speaking Across Tiers: Gender, Communication, and Institutional Incentives in Mixed-Member Elections

Political candidates, and women in particular, face a variety of challenges in how they present themselves to voters. Research shows that women are expected to focus on specific issues, avoid overly aggressive tones, and communicate in ways that align with gendered expectations. Yet we know less about how these constraints interact with electoral incentives and how they shape what politicians choose to emphasize or engage with in digital environments. Mixed-member electoral systems offer a useful context for studying this interaction because they embed majoritarian and proportional rules within the same election, allowing comparisons across electoral tiers while holding constant broader national conditions. The institutional contrast matters: majoritarian candidates rely on personalistic, visibility-driven, broad constituency representation, whereas proportional candidates operate within a more party-centered, group-representation logic. For women, these incentives can pull in opposite directions—between communicating as representatives of women as a group and presenting themselves as broadly appealing, region-oriented candidates. How this tension plays out remains underexplored. We theorize that the group-representational logic associated with proportional representation is only partially retained in mixed systems, while majoritarian tiers amplify personalistic communication. These institutional pressures intersect with gendered expectations: women may emphasize collective representation such as women’s rights on PR lists, while majoritarian competition pushes them toward localized, constituency-focused appeals, including highlighting “my region” or district-specific concerns. We also assess whether candidates differ in how they engage with citizens online, including whether they escalate or de-escalate contentious interactions, and whether majoritarian candidates, whose electoral fortunes depend more on individual visibility, interact more frequently with the public. We analyze 766,945 tweets from 4,620 candidates competing in elections between 2021 and 2022 across Germany, Italy, Japan, Nepal, and the Philippines. These countries vary widely in gender norms, party strength, and democratic development, allowing us to test when institutional incentives reinforce or dilute gendered communication patterns. Using computational text analysis and supervised machine learning, we examine topic emphasis, references to regional identity, engagement strategies, across gender and electoral tiers. Our findings show how electoral rules and gender norms jointly shape digital political communication, illuminating how candidates navigate competing pressures for visibility, representation, and electability.