Search for contacts, projects,
courses and publications

Referendum Forecasts, Turnout and Ballot Results



Funk P. F.


Savio G.



This project investigates whether ballot forecasts on initiatives and referendums affect participation decisions of potential voters. Knowing whether and how pre-polls affect voters’ participation decisions is crucial: if pre-poll results affect certain individuals’ participation decisions, then pre-polls themselves may have an impact on the voting result. Put differently, not displaying any pre-poll results may lead to different electoral outcomes.

As long as pre-polls correctly display voting intensions, potential voters can rationally decide whether to participate based on all available information. For instance, if a vote is predicted to pass or not pass by a large amount, some citizens may not find it worthwhile to incur the costs of turning out. Releasing pre-poll results may be beneficial, as it safes voting costs for citizens whose vote very unlikely would have had an impact. However, it is also possible that pre-polls incorrectly display voter preferences, and this leads to voting mistakes. As such, methods that correctly elicit voting intentions are less likely to lead to voting mistakes.

Understanding whether and how pre-polls affect turnout decisions is of obvious importance for a direct democracy such as Switzerland. However, other European countries have conducted referenda on very important issues (referendum on Scottish independence), where pre-polls results were available. As such, this research question is of much broader importance than the Swiss context alone.

The first part of the project investigates the turnout effects of pre-polls results, using Swiss data on all ballots held since 1998. Since the gfs ( releases results from telephone surveys held shortly before the votes, we can investigate how the ballot forecast influences subsequent turnout. The second part of the project compares different survey techniques in their accuracy to elicit voter preferences. Understanding which survey methods are best in eliciting true voter preferences is crucial given that surveys are likely to have an impact on voter decisions and hence electoral outcomes.

Additional information

Start date
End date
36 Months
Funding sources
Swiss National Science Foundation / Project Funding / Humanities and social sciences (Division I)