Votendi’s new experimental weighted polling index places the SNS/Vučić-led political family at 43.6% and the student-backed list family at 41.1% as of 22 October 2026, a 2.5-percentage-point advantage. The result combines ten eligible national surveys from four polling organisations. It is an analytical index, not a new survey, election forecast or estimate of seats.
With Serbia’s parliamentary election scheduled for 25 October, the index offers a consolidated view of available polling while revealing a persistent problem: individual pollsters still disagree on which side leads.
Votendi’s Serbia election hub now displays the weighted index alongside observed ranges, threshold watch, individual poll comparisons and the methodology.
Votendi’s experimental index: SNS 43.6%, Student List 41.1%
| Political family used for comparison | Weighted index |
|---|---|
| SNS / Vučić-led configurations | 43.6% |
| Student-list configurations | 41.1% |
| Difference | SNS +2.5 percentage points |
Votendi experimental weighted index, calculated as of 22 October 2026 from ten eligible surveys by four polling organisations. Figures rounded to one decimal place; not a forecast.
The index groups related names used across surveys, including “Aleksandar Vučić – Ujedinjena Srbija” and SNS configurations on one side, and “Studentska lista”, “Student List” and “Vladan Đokić – Studenti pobeđuju” on the other.
These are analytical families, not claims that every poll tested identical registered ballot lists. Original labels remain available in individual source records.
The latest two surveys tell different stories
The most recently published figures do not line up neatly.
Nacija TV / Nacija portal’s poll, published 21 October, put the Vučić-led configuration at 42.9% and the student-backed configuration at 42.0%, a 0.9-point government-side lead. Its fieldwork ran from 16 to 18 October.
By contrast, NSPM’s poll, published 20 October, reported 44.5% for the Student List and 42.7% for the Vučić-led list, a 1.8-point student lead. Its interviews were conducted between 25 September and 5 October, meaning the newer publication date should not be confused with newer fieldwork.
The contrast grows when earlier surveys are considered.
A Faktor Plus survey fielded in September put the government-side configuration at 48.6% and the Student List at 37.1%.
A CRTA / Stanford Democracy Action Lab survey conducted in June reported 35.7% for SNS and 44.9% for the Student List among its reported voting-intention responses.
These results concern different fieldwork periods and potentially different voter bases. Their divergent leads are not evidence of a sudden electoral swing.
How Votendi calculates the weighted index
The experimental model includes published national surveys that report both broad political families.
A survey’s influence declines with age, using a 30-day half-life calculated from the end of fieldwork when known, or publication date when fieldwork details are unavailable.
The model also scales weights by the square root of effective sample size, with upper and lower limits, and reduces the weight of records missing key metadata.
For the October Nacija online poll, the model uses its disclosed effective sample size of 599, rather than the nominal sample of 1,204.
To avoid allowing organisations that publish more frequently to dominate automatically, each pollster’s total influence is limited to the strength of its strongest eligible survey and distributed across its included waves.
Those calculations produce an index of 43.6% to 41.1% on the data available on 22 October.
The result is descriptive, not predictive. It does not estimate pollster house effects, adjust all questionnaires onto an identical voter denominator, calculate statistical confidence intervals or project parliamentary seats.
The historical line connecting index snapshots is a visual guide to published information, not a measurement of voter opinion every day.
Why the index does not settle the election
The 2.5-point difference is the output of a stated weighting method.
It is not proof that SNS has a statistically secure lead, especially when individual surveys vary widely in design, fieldwork dates, electoral-list names and treatment of undecided voters.
Combining conflicting measurements cannot remove uncertainty in the underlying data.
Votendi will therefore continue publishing each eligible poll with its original source, dates, sample information and reported list labels.
Readers can compare the figures directly and inspect the model’s included surveys and relative weights through the experimental index and methodology.
For additional context, see Votendi’s earlier comparison of Serbia’s diverging pollsters and analysis of the October NSPM release.
Data cutoff: 22 October 2026. Weighted index figures are Votendi calculations from recorded source surveys. No proprietary Serbia election forecast is offered.
