Prepared by: Etrit Rexhepi
Yesterday, on October 12, local elections were held in Kosovo. Immediately after the end of the voting process, three television stations published the results of exit polls, which predicted the winners in several municipalities of the country. The exit poll measurements were carried out by UBO Consulting, PIPOS and Albanian Post, and published on television stations RTV Dukagjini, "Clan of Kosovo”And“T7".
In all these predictions, it was stated that in the Municipality of South Mitrovica, Arian Tahiri, the candidate of the Democratic Party of Kosovo (PDK), has managed to receive the majority of votes to receive the mandate of mayor of this municipality in the first round (see here, here and here). However, after counting the votes in this municipality, according to the publication on the website of the Central Election Commission, the results showed a difference of 28 votes between Tahiri (47.89%) and Faton Peci (47.80%) of the Vetëvendosje Movement. This result The current situation means that there will be a runoff (second round) between these two candidates for mayor of South Mitrovica.
The results from these elections showed that the predictions for South Mitrovica turned out to be incorrect. But what are exit polls and what could be their impact on public opinion?
According to the first chapter of the book "Exit Polls – Surveying the American Electorate, 1972-2010" (Albanian, "Exit Polls - American Electorate Polls, 1972–2010"), compiled by two authors and researchers of Political Science in the United States of America (USA), Samuel J. Best and Brian S. Krueger, it is said that exit polls are surveys conducted with voters immediately after they leave the polling stations.
They are used by the media to predict the winner before the vote count is complete, to analyze how different social groups voted, to understand which issues influenced voters' decision-making, and to interpret political results after elections.
However, the results of these surveys often appear briefly in the media and then disappear from public access, remaining archived in universities.
How exit polls work
The book says the process requires up to two years of preparation, including selecting polling stations, representing the entire country, interviewing voters exiting the center, depending on the designation, telephone surveys for early voters or by mail, questionnaire design, with 25–40 questions on demographics, political orientation, economics, and statistical analysis and weighting to correct for nonresponse and sampling errors.
History
The book shows that exit polls were born in the 1960s–70s at CBS News by Warren Mitofsky and Murray Edelman, to improve election predictions.
During the 80s, all the major television networks (CBS, ABC, NBC, LA Times) conducted their own polls. They began to be used for early projections, which caused controversy, as they were often broadcast before the polls closed.
After 1988, due to high costs, the networks merged into a joint consortium – Voter Research and Surveys (VRS), which later became Voter News Service (VNS).
Error cases
The book is written by American authors, so the examples are taken from US elections. In 1990–1992, VRS had technical problems and errors in statistical weightings. In the 1992 presidential election, preliminary exit poll results overestimated the vote for Bill Clinton and underestimated that for George HW Bush, creating a false perception of a major Democratic victory.
The early release of this data influenced public opinion and media reporting, leading to premature celebrations and political reactions.
The most infamous case in the book is the Bush-Gore election. The television networks, relying on data from VNS, mistakenly declared Al Gore the winner in Florida, then retracted the statement, and then declared Bush the winner, only to retract it again later.
This event caused public chaos, loss of trust in the media, and a sense of manipulation among American opinion.
strengths
Exit polls measure real voters immediately after voting. In the US, they allow self-administration of the selected ones while preserving their anonymity, which reduces social pressure. They also include very large samples (the book mentions figures of 10–18 thousand people), which allows for detailed analysis by age group, race, religion, etc.
Limitations and public impact of potential errors
Although more advanced, exit polls are subject to four typical sources of error. They are: sampling error, coverage error (omission of some voters, especially early voters), non-response (voters who refuse to participate), and measurement error (misunderstanding of questions or political sensitivities).
These errors, even when small, carry great weight in public opinion because they can create false perceptions about the winner, can influence the behavior of other voters when published early, undermine trust in the media and the electoral process, and are often used politically to fuel theories of manipulation or to delegitimize the result.
Local elections 2025 in Kosovo
The latest case in Kosovo has sparked debates about the reliability of these measurements. In fact, the acting Prime Minister and leader of the Vetëvendosje Movement, Albin Kurti, today (October 13) has called for an investigation into the exit poll reports published on October 12 for South Mitrovica, describing them as “flagrant interference” and not simply as “misleading”. In a press conference, he said that the predictions published by several organizations and television stations, which showed the victory of the PDK candidate before the closing of the polling stations and with a difference of 16 percentage points from the real results, constitute “premeditated engineering” and “an infringement of democracy and the integrity of the state”. Kurti called on the security and justice bodies to investigate the case, emphasizing that propaganda had been carried out for Mitrovica since the first hours of election day (see here).
While Baton Haxhiu, author of one of the exit polls, said that the error in these measurements in South Mitrovica was human, but emphasized that the reaction of the acting Prime Minister Albin Kurti later turned it into a political battle. He said that the problem does not lie in the survey, but in the "climate of fear that the government has created, where voters feel watched and insecure to freely express their opinion". According to him, in Mitrovica, the free will of the citizens was not measured, but their fear. Haxhiu concluded that Kurti used this error to turn it into a moral victory (see here).
So, the audience should keep in mind that exit polls are surveys conducted immediately after the vote to predict winners, analyze the preferences of social groups, and understand the issues that influence voters’ decision-making. However, they are not infallible. Errors can arise from sample selection, limited coverage, voter non-response, or mismeasurements in questionnaires. Published exit polls (especially those published early) can influence public opinion, create false perceptions, and undermine trust in the electoral process. The recent case in Kosovo, where exit polls for South Mitrovica did not match the actual results, shows the importance of using these measurements carefully.