Update
University of Belgrade rector infected with Pegasus, organiz...
Citizen Lab confirmed a zero-click Pegasus infection on rector Vladan Đokić’s phone, while Rad...
2,500
observers in CRTA EOM
500
polling stations in the sample
2×2
observers per station: two in the morning shift, two in the afternoon
What CRTA does on Election Day
Process
SBO · statistically based observation
Observers are at the polling station from preparations for opening until the results are announced. They record the opening, voter identification, secrecy of the vote, the presence of list representatives, closing, the count and the results protocol.
Result
PVT · parallel vote tabulation
At the end of the day, observers record the officially announced figures from their polling station. Aggregated across the sample, these give a projection of the result and turnout, and a check on the official outcome.
Since their first application in the late 1980s, PVTs have been conducted by citizen observer organisations in every region of the world. This globally recognised methodology allows domestic observers to assess, on the basis of verifiable evidence, both how Election Day is conducted and whether the officially reported results reflect the ballots cast.
What CRTA does not do
Not an exit poll. We do not ask citizens how they voted.
Does not declare a winner. That is the role of the electoral commission.
Does not read the numbers without the process. A flawed process undermines even an accurate count.
Three numbers to watch
Key statistical references for anyone following CRTA's findings on Election Day.
1
Response rate
How many sampled polling stations have reported their data. It must be close to 100%, otherwise the sample is not representative and there is no projection. This is checked first.
2
Margin of error
The PVT gives a range, not an exact percentage, at a 95% confidence level. The margin measures statistical precision, not a calculation error.
3
Turnout estimate
Also a range. During the day it can point to irregularities, for example a jump in turnout larger than would be possible under lawful conditions.
Election Day, step by step
-
1Before Election Day: the sample is drawn, and observers are recruited, trained and deployed
Statisticians design an independent, random and representative sample of 500 of Serbia's 8,210 polling stations - large enough to support a statistical claim about the whole country. The sample is stratified by three categories - region, polling-station size and the political profile of the area - each divided into strata (1.1 to 3.5). CRTA will report results once data from all three strata have been collected.
-
2Throughout the day: CRTA reports on the turnout projection, process and atmosphere
Using standardised forms, observers report in real time on turnout, the conduct of key procedures, the atmosphere and any incidents. Irregularities are logged to a legal standard, so they can support complaints.
What CRTA will publish on Election Day10:00Turnout projection13:00Turnout projection16:00Turnout projection19:00Turnout projectionAfter 21:00Results projectionThe results projection follows as soon as the sampled polling stations in all three strata have finished counting.
-
3After polls close: observers use PVT to collect the results
After the count, observers record the official results announced by the polling board at each sampled polling station and send them to CRTA HQ, where each report is verified. These results form the basis of the PVT.
-
4Election night: CRTA announces its projection of the results, aggregated from the PVT data
When CRTA can announce its results depends on the pace of the count at polling stations. The PVT estimate can be made only once polling boards at the sampled stations have counted the ballots and determined the results, so any delays in counting will also delay CRTA's announcement.
The logic of the checkThe PVT is not meant to announce the result first, but to check the official one once it is published.Official result within the range: consistent with the observation, no signs of manipulation in the count.
Official result outside the range: a serious signal that requires an explanation.
-
5The day after: CRTA publishes its Preliminary Statement on Election Day
The statement will set out CRTA's assessment of the quality of Election Day and the integrity of the outcome.
Follow the mission on 25 October
From its Belgrade headquarters, CRTA will publish findings on turnout and irregularities during the day, and on the results once the PVT is complete. Everything the mission says publicly is traceable to verified data.
Elections 2026 on CRTA.Plus →CRTA's Election Observation Mission applies the PVT methodology, combined with long-term observation, in line with the Declaration of Global Principles for Non-Partisan Election Observation and Monitoring by Citizen Organizations.
