A gift to democracy? The campaign learning effects of targeted online political ads in the Dutch 2023 election campaign

Scientific work on data driven campaigning often focuses on mapping its threats, such as voter manipulation, but there is little attention to the potentially beneficial aspects. Arguably, potential benefits are the informative function (targeted) ads can have: they can provide voters with relevant information about parties and issues in an efficient manner, in particular voters who are difficult to reach with traditional media (Zuiderveen Borgesius et al. 2018). However, we lack knowledge on whether online ads really inform voters during election campaigns and whether targeting contributes to this. Investigating this is theoretically important since the scholarly focus is often on persuasion while the informative potential of online campaigns is understudied. We expect knowledge of political candidates and viewpoints to increase, especially when voters receive ads that fit their political preferences (targeted ads). This increase in knowledge is more likely to be observed concerning newer political parties and candidates than established ones. To the best of our knowledge, this is the first longitudinal study on the campaign learning effects of online (targeted) political ads. Our paper focuses on the 2023 Dutch parliamentary elections. This election serves as a most-likely case for voters to learn from political ads due to a combination of unique factors. These are an unprecedented number of new (relatively unknown) party leaders, two new parties that are likely to attract many voters, but with unfamiliar standpoints and candidates compared to established parties, and, lastly, the presence of a large group of “floating’ voters who are likely to make up their minds in the final weeks, or days, before the elections (van der Schelde 2023). Data was collected through a panel study in the run-up to the Dutch general election in November 2023 together with polling agency I&O Research. We also deployed the mobile Experience Sampling Method (mESM) with data donations among a subsample of these participants. This consisted of an intensive three-week study during which they were asked to send screenshots of the political ads they saw on social media (e.g., FB, Instagram, YouTube, TikTok) and while browsing (e.g., banners on websites). Ads will be coded on party and issue positions. This study links mESM data and survey data. After each upload respondents were asked whether they learned anything about the candidates or the viewpoints of the party that posted the ad. Before, during, and after this period campaign-specific political knowledge was measured multiple times in the overarching panel survey. Candidate-specific knowledge was measured through the ability to correctly identify candidates and their party affiliation. To measure issue-specific knowledge voters were asked to correctly position their preferred party on a number of issues (van der Meer et al. 2016). Results will inform us about the differential impact of microtargeting on democracy.