ENHU

New open course led by colleague Gábor Tóka

16/03/2026

Election Prospects in Light of the Polls

The language of the seminars is Hungarian. For more information, please visit the Hungarian language page: https://archivum.org/hu/news/new-open-course-led-by-colleague-gabor-toka

The open course, organized and led by Archivum colleague Gábor Tóka will constitute of three seminars.

1. Which opinion poll can we trust?

March 20, 2026, 17:30-19:00

Why are there so many polls, who pays for them, is there a reasonable explanation for the discrepancies in the results, and which ones should we believe? How accurate were they in the previous Hungarian elections, why were they wrong, and is domestic polling worse than the global average? How could they be improved?

Our discussants will be practicing pollsters who perhaps know the most right now about the difficulties currently facing in-person, telephone, and online surveys in Hungary, and what these surveys reveal about public sentiment.

Roundtable participants:

Ábel Bojár (Europion)
Mária Némethné Czinkóczi (IMAS International Hungary)
Andrea Szabó (ELTE Center for Social Sciences)
Gábor Tóka

2. Who stands to gain the most in April? – Seat allocation and the parliamentary horseshoe

Aprilis 2, 2026, 17:30-19:00

Why is it harder to predict results in individual districts than national vote shares? What methods can be used, and how accurate are they? How should a seat calculator be designed, and why do they not all show exactly the same result? Can we trust any of them at all, and what is the margin of error in their estimates? How can averaging different polls go wrong?

Today’s participants are leading expert bloggers who have created what are perhaps the most sophisticated seat calculators for this year’s Hungarian election.

Roundtable participants:

Bálint Horváth (voksadat.hu)
Tamás Kovalcsik (Partizán Adatpont)
Ádám Sanyó (Taktikai Szavazás blog)
Viktor Tisza (SzázKilencvenKilenc blog)
Gábor Tóka (Vox Populi election guide)

3. Megjósolható-e a választás kimenetele?

Aprilis 9, 2026, 17:30-19:00

Can election outcomes be predicted based on social media analysis, economic trends, the results of midterm local elections, or betting markets? Can we use artificial intelligence to fill out surveys the way people of a given demographic would in a similar situation? Can we entrust the work of interviewers to a machine capable of learning, and will this make the interviews better? How do pollsters tackling such serious questions attempt to determine what election results can be expected next Sunday?

Today’s participants are researchers who have experimented with—or are currently experimenting with—innovative methods, virtually unknown to the general public, in order to better understand the domestic political and social climate.

Roundtable participants: