Dimitrijevic on RIK's first 203 polling stations

Republic Electoral Commission (RIK) chairman Vladimir Dimitrijevic told the Tanjug and Beta news agencies on Friday, October 30, that data from the first 203 processed polling stations, which some media published on election night, October 25, were not RIK’s preliminary results. At the same time, he said the figures were official data from the commission’s system.

The two agencies carried the same statement with a different emphasis. Tanjug writes that Dimitrijevic confirmed the Tanjug agency and television channel published accurate data on election night, taken from the system at rik.parlament.gov.rs rather than from a press conference. Beta says Dimitrijevic stated that RIK did not announce results from the first 203 polling stations, and that the system simply recorded them as the minutes arrived. Both reports come from the same statement, and the difference lies in which part each agency puts first.

How did the data reach the RIK system?

Dimitrijevic described the procedure from the moment polls close. Tanjug quotes him:

“After the polling stations close, which happens at 8 p.m., the counting of votes begins. When the polling boards count the votes, establish the voting results, and fill in and sign the polling board’s minutes, the chairman and other members of the board go to their local electoral commission to hand over the materials, among other things the minutes of the polling board’s work. On that occasion the minutes are scanned, and statistical processing is carried out by representatives of the Republic Statistical Office, so they enter it into this system of ours.”

The RIK chairman said the minutes are entered in the order they arrive, so the first entries began as soon as the polling boards finished their work:

“And verification is carried out, starting from the first polling station entered in Serbia, and then in order as the results come in. That generally happens when a polling board finishes, somewhere around 9 or 9:15 p.m., and the entry begins.”

Dimitrijevic added for Tanjug that he is not sure whether those 203 stations were entered around 9:15 or 9:30 p.m., but that everything was visible on the screens at RIK headquarters, in the hall outside the room where the commission sits. Speaking to Beta, he said it is possible that the first 203 minutes immediately gave the picture the media reported:

“When the polling stations close, the polling boards establish the voting results, and that material, together with the minutes, is taken to the municipal electoral commission. The minutes are scanned and entered into the system, and those results are immediately published on the RIK website. It is possible that those first 203 minutes, entered at random from all over Serbia, pointed to a result that some media reported right away.”

Beta adds that, according to its report, television stations citing RIK published figures by list at around 9:20 p.m. on Sunday.

Why are 203 polling stations not a representative sample?

According to Dimitrijevic, the first processed minutes came from smaller, mostly rural polling stations, because their boards finish faster. He told Tanjug:

“So these are polling stations with 100 voters, 200, 300, 400 voters. They are processed faster, people bring the materials to the local electoral commission faster, and that is how they get recorded. That is the essence of it. Now, which polling stations those are, they are random ones, all over Serbia, I mean, that is how it is.”

Asked whether the data Tanjug published was official, he answered that it was, but that it was not a reference point. He also compared the number of stations with the total: 203 out of 8,392 is a little under two and a half percent.

“They are, that is right, they are official data, but that is not any kind of reference figure. For example, 200 polling stations against the 8,392 there are in total. If you remember, in some earlier election cycles, before this kind of data processing system was introduced, RIK came out with an announcement only when roughly 30 percent of polling stations had been entered.

And that was in some earlier election cycles, back in 2016, 2017, 2020, 2022, and so on. Then this system was introduced where it is published publicly, literally starting from the first station processed, so that it can physically be seen.”

Dimitrijevic also said the data in the system changes quickly. Tanjug paraphrases him as saying this happened within fifteen minutes or so, while Beta gives a verbatim quote:

“RIK did not publish the results from those first 203 processed polling stations. What it was is a system available to the public and journalists, and that data changed from hour to hour, as the minutes from the other polling stations were entered.”

When did RIK announce its first preliminary results?

The RIK chairman maintains that the commission did not convene on election night and did not adopt any preliminary results. According to Tanjug and RTV, he said:

“We did not convene or adopt any, that is, preliminary results. We only adopted them 24 hours after the polling stations closed, on Monday after 8 p.m., when almost 98.5 percent had been entered.”

Beta reports that RIK announced its first preliminary results on Monday evening, 24 hours after voting ended. The 98.5 percent figure in Dimitrijevic’s statement refers to entered polling stations, and it appears only in the Tanjug and RTV report, not in Beta’s.

What does Tanjug say about the reactions to its report?

In the same piece, Tanjug says the agency and the television channel were accused in recent days of tendentiously publishing inaccurate data while citing RIK. According to Tanjug, the newsrooms of K1 television and the Tanjug agency received threats on Monday evening, which were reported to the competent institutions, and the European Alliance of News Agencies (EANA) condemned the threats and intimidation of employees. The same source says K1 and Tanjug halted live programming on Tuesday while agency staff worked from home, and that threats were also made against editor-in-chief Jovana Joksimovic.

Dimitrijevic told Tanjug that the minutes were in order, that they were scanned and posted on the RIK website, in the place where records are kept by polling station:

“It is not possible for some big polling stations to be entered, because simply it takes time to process them. So these are some smaller polling stations, which were processed at that moment. The minutes were in order, statistically as well, and afterwards they were scanned and posted on our website, where the records are kept by polling station.”