Swiss Users’ Search Practices on Political Referendum Topics

The next presenter in this AoIR 2022 session is my current University of Zürich colleague Sina Blassnig, who shifts our focus to the users of social media platforms. They need political knowledge to make rational decisions, but this is difficult in today’s high-choice informational environments; one key source for such information, of course, are search engines, but research on their role with regard to political issues and referenda remains very limited. The current study explores this in the Swiss content, examining how often Swiss citizens search for information on upcoming referenda. Generally, such search practices may be related to demographic factors.

The quality of search results is also related to the algorithms employed by search engines, and this has given rise to the myth of the ‘filter bubble’; in reality, differences in the search results people receive will have much more to do with their use of these platforms, and especially the formulation of search queries. The present project explores these issues in relation to a number of specific recent referenda in Switzerland, as related to COVID-19 policy, a care initiative, and the choice of judges by lot.

It obtained Google search histories from users through data donations, and enriched the data with further survey responses about sociodemographic characteristics and political positioning; this is also difficult because of the genuine concerns people have about giving up their search histories, however. The project also asked participants whether they voted in the three referenda, and what they voted for; and it explore whether their search histories showed any interests related to the referenda topics. The final sample contained 128 participants.

Of the 150,000 search terms used by the participants, only 90 were actually related to the referenda; 72% of these related to the COVID-19 referendum alone. Those who searched for issues related to the referenda were more likely to be young, male, and politically interested; the more important they perceived an issue to be, on the other hand, the less likely they were to search for the topic.

Most of the search terms were fairly neutral; this was especially true for non-voters. Pro Websites showing up in searches were exclusively visited by voters in favour; contra Websites exclusively by voters against the issues central to each referendum. These results are not necessarily representative and come from the referenda-friendly nation of Switzerland, of course, so they may not translate to countries where referenda are considerably more rare.