The next speaker at the DigiWorld network ECREA preconference on election campaigning in Brno is Kajsa Falasca, whose focus is on TikTok engagement ahead of the upcoming Swedish election. TikTok analysis is complicated, and requires an attention to multimodal content; this study employed an approach called Multimodal Narrative Mapping, using a methodological approach which automatically scrolls through videos and analyses the content of the videos displayed.
In this sense, then, TikTok changes both the process of campaigning and the object of analysis: content served to users extends beyond the accounts that users themselves follow, and is also determined by what their friends follow, and a range of other parameters determining the curation of the For You feed. Such content, then, is highly multimodal, and political leaders have generally responded to this by personalising their communication styles.
This project focussed on nine political leaders, and analysed their TikTok videos using an AI-driven approach; the AI coding was in turn verified by three human coders for a subset of the total dataset. This separates three dimensions: what leaders choose to communicate; how audiences engage with this; and what types of content overperform relative to the frequency of content.
Migration and energy security emerged as the most widely shared policy fields across all leaders. However, some leaders employed an issue specialist narrative profile, focussing strongly on only some of these issues; while others – especially those with ambitions for the Prime Ministership – broadly covered most issues; this did not manifest in matching engagement patterns, however, where issues such as equality and healthcare were most prominent.
