A new Ipsos Strategic Marketing survey published on 22 October puts Serbia’s Aleksandar Vučić – Ujedinjena Srbija electoral list at 47.4% and the Student List at 40.0%, a 7.4-percentage-point gap. Votendi has also calculated how the published figures translate mechanically into Serbia’s 250 parliamentary seats under the electoral formula. That seat calculation is an illustrative scenario, not a forecast of Sunday’s election.
The survey was conducted from 20 to 22 October 2026, three days before the 25 October parliamentary election. Ipsos interviewed 3,624 adults in Serbia, excluding Kosovo and residents abroad. According to the official Ipsos release, 90% of interviews were conducted by telephone and 10% online.
Ipsos discloses that the research was commissioned by the electoral list Aleksandar Vučić – Ujedinjena Srbija. That funding information is essential context when comparing survey results, though it does not by itself establish their accuracy or inaccuracy.
Ipsos poll: Vučić 47.4%, Student List 40.0%
| Electoral list | Estimated share | Published 95% interval |
|---|---|---|
| Aleksandar Vučić – Ujedinjena Srbija | 47.4% | 45.3–49.5% |
| Studentska lista – Studenti pobeđuju | 40.0% | 37.9–42.1% |
| Ivica Dačić – Faktor stabilnosti | 3.2% | 2.5–3.9% |
| Autentična desnica – Miloš Jovanović | 1.7% | 1.1–2.3% |
| Alliance of Vojvodina Hungarians | 1.2% | 0.7–1.7% |
| Evropska Srbija – Tepić / Ponoš | 0.8% | 0.4–1.2% |
Ipsos estimates turnout at 64.7%, with a stated interval of 62.9% to 66.5%. Vote shares refer to respondents reporting a voting choice, while the estimates also allow for an assumed 2.5% of invalid ballots. The 18 lists’ published percentages therefore add to 97.5%, not 100%.
Dačić’s list, at 3.2%, is particularly sensitive to the 3% ordinary-list threshold. Its reported interval runs from 2.5% to 3.9%. Under the point estimates, the other non-minority lists shown here fall below 3%.
What a mechanical 250-seat allocation shows
Serbia elects 250 MPs in one national constituency. The ordinary electoral threshold is 3%, but formally recognised national-minority lists have an exemption. When such minority lists receive under 3%, their D’Hondt quotients are increased by 35%. Votendi applied those rules to Ipsos’s exact published list shares, treating them as fixed vote totals solely for this arithmetic exercise.
| Electoral list | Ipsos share | Calculated seats |
|---|---|---|
| Aleksandar Vučić – Ujedinjena Srbija | 47.4% | 127 |
| Studentska lista – Studenti pobeđuju | 40.0% | 107 |
| Ivica Dačić – Faktor stabilnosti | 3.2% | 8 |
| Alliance of Vojvodina Hungarians | 1.2% | 4 |
| Rasim Ljajić – Ljudski | 0.5% | 1 |
| Usame Zukorlić – Pokret za ujedinjenje | 0.3% | 1 |
| SDA Sandžaka | 0.3% | 1 |
| Albanian Coalition | 0.3% | 1 |
| Total | 250 |
The arithmetic gives 127 seats to the Vučić list, 107 to the Student List, eight to Dačić’s list and eight combined to five minority lists. A majority in the 250-seat chamber requires 126 seats. This describes the output of one formula with one poll’s point estimates; it says nothing about the probability of any side obtaining that majority on election day.
The allocation is sensitive to even modest changes. Moving one percentage point from the Vučić list to the Student List while holding every other Ipsos percentage unchanged changes the same calculation from 127–107 to 124–110 seats for those two lists. This is a sensitivity demonstration, not a prediction of voter movement.
Why Ipsos and NSPM show different races
The most recent NSPM release, published 20 October, reported 44.5% for the Student List and 42.7% for the Vučić list: a 1.8-point Student List lead. That contrasts with the 7.4-point Vučić-list lead in Ipsos’s new survey, a 9.2-point difference in the two lead margins.
Those figures should not be interpreted as a sudden nine-point swing. NSPM’s telephone fieldwork took place 25 September–5 October, whereas Ipsos surveyed respondents on 20–22 October. Sampling, fieldwork timing, questionnaire design, handling of undecided voters and commissioning arrangements differ. The two surveys are not directly interchangeable measurements made on the same day.
Another Nacija TV poll, published 21 October, reported 42.9% for the Vučić list and 42.0% for the Student List. The published poll figures therefore continue to diverge even during the final week of the campaign.
Votendi’s separately labelled experimental weighted index, reported earlier on 22 October at 43.6% versus 41.1%, is an analytical combination of eligible previous polls, not an Ipsos result. Readers should not mistake that earlier index snapshot for an automatically recalculated value including today’s Ipsos survey.
How to interpret the last polls before 25 October
A single poll provides an estimate for a specified population and time period, not a certified count of votes. Ipsos itself notes that some voters make their choice in the final days. Its published confidence intervals are also not a guarantee about the final results; survey bias, turnout assumptions and differences between polling organisations matter.
For the full record of polling data, list definitions and methodology, see the Votendi Serbia election hub and our election guide. For the formula governing parliamentary seats, see the Republic Electoral Commission’s election-system explanation and Article 140 of the electoral law.
Data and methodology note: Ipsos, published 22 October 2026, fieldwork 20–22 October, n=3,624. Independent seat exercise: Votendi, 22 October; 18 reported electoral-list shares, 3% standard threshold, minority exemptions and 35% minority quotient adjustment, 250 D’Hondt quotients. This fixed-input scenario is neither a poll average nor an election forecast.
