How Vucic dropped from 56 to 48% and Radulovic rose from zero to six
Did CeSID’s motto, 1 + 1 = 2, hold true on election night? The preliminary results caused even more confusion in an already tense election atmosphere, marked by claims of numerous irregularities and possible electoral fraud.
In under two hours, Aleksandar Vucic and the Serbian Progressive Party (SNS) “fell” from a historic landslide to a repeat of their results from the last election. The Enough Is Enough (Dosta je bilo) – Sasa Radulovic movement went from no mention to over the parliament threshold, while the Alliance of Vojvodina Hungarians (SVM) was at one point attributed 32% of the ballot in Vojvodina.
Election night saw preliminary results released by CeSID, Ipsos and CRTA (observing the vote for the first time). All had SNS taking in excess of 50% of the ballot, more or less, in their first projections.
As the Republic Electoral Commission (RIK) made no statements – and considering that Aleksandar Vucic had said during the campaign that the SNS goal was to have at least one vote more than all the other parties put together, these about-faces in the projections sparked suspicion among the public that the watchdogs we had for years trusted to guard the electoral will of the people were party to dishonorable acts.
CeSID mere nuisance
“After 20 years, CeSID was the only nuisance on election night,” the organization’s former program director Zoran Lucic told Insajder.net.
He believes that the entire situation arose because the work was done by unskilled and undereducated people who don’t have sufficient knowledge of mathematics or elections, “and they thought they had mastered it all.”
“If you watch a plane fly from afar, that doesn’t mean you can fly it too,” says Lucic.
“CeSID didn’t do its job professionally. What I see is that they had a bad sample, it wasn’t stratified. What happened in Vojvodina— for you to say that SVM has 30%, that’s unacceptable,” he goes on.
The first preliminary results released by CeSID assigned 56% of all votes cast to SNS.
Many found it odd that the Enough Is Enough movement wasn’t even mentioned.
In the end, CeSID’s results have SNS below the famed 50% – specifically, at 48.88.
What does CeSID say?
The explanation of Bojan Klacar of CeSID (Center for Free Elections and Democracy) is that this was the first time that data were presented almost in real time, as they came in, i.e. every 15 minutes. Still, Klačar admits that there may not have been a sufficiently clear explanation of what type of preliminary results were being released.
“First and foremost, we released the results as the sample grew. We had two types of polling stations, rural and smaller polling stations, and city stations. That means that we would first get results from places where there are 100 or 200 voters. The sample was first filled from polling stations from central Serbia, then Vojvodina, and only then Belgrade. We made our first projection based on a sample filled with around 17% of the material, and essentially most of it was from central Serbia. The changes followed because data came in from big polling stations. Enough Is Enough has votes in Vojvodina and in Belgrade. That’s when the results started to change in their favor,” Klačar explained for Insajder.net.
As proof that CeSID’s sample was good, Klačar pointed out that the results it released around midnight match those of RIK.
“We used the same method in this election that CeSID has used since 1997. The only difference is that we published the results live every 15 minutes. Before, we would call a conference say at 9:30 p.m. and another when we would give the tally of the seats, and that was that. Now we presented the data to the public as they came in. Had we released the results the same way in earlier years, we would have had the same situation and the same changes.”
The media rush too
Srđan Bogosavljević of Ipsos Strategic Marketing explains the change in the percentage of the votes won by SNS, which dropped from 55 to 48% in Ipsos’ estimate of the results in a short time, as the consequence of quicker inflow of results from smaller polling stations and slow counting due to the high number of tickets at both the national and the provincial level.
“Also, it’s important to the media to have any kind of result as soon as possible,” says Bogosavljevic.
What Ipsos considers relevant information are the figures from scoring 40% of all parts of the sample – that is when, according to Bogosavljevic, the figures stabilized, and it turned out that those results matched the ones that RIK gave early on April 25.
“The Serbian Progressive Party was under 50% as early as the verification of 20% of the sample, according to Ipsos’ data. The problem is that you didn’t know how it would go with the parties around the threshold.”
Bogosavljević noted that Ipsos couldn’t understand why there was such a difference between their and CeSID’s data.
CRTA: We didn’t rush
The Center for Research, Transparency and Accountability (CRTA), a group of activists and reporters monitoring the electoral process for the first time, also participated in the release of initial preliminary results during election night. CRTA says it didn’t want to come out with any figures before getting 65% of the strata, which means covering every segment of the sample from all polling stations, and until the percentages stabilized, i.e. when they stopped changing suddenly as information came in.
CRTA was late in publishing the first tally of the results for that very reason – their focus was on quality and relevance.