Original article (in Montenegrin) was published on 10/03/2023
At the mention of CSI, most people think of the crime series about the forensic team from Miami, its main character Horatio Caine and his iconic manner of putting the sunglasses on at the moment when he delivers one of his one-liners.
However, in the last few election cycles, CSI has been associated with something completely different from forensics and sunny Miami – public opinion polls and surveys exclusively related to the elections in Montenegro.
Since public opinion surveys always attract the attention of the public, especially in the run-up to the elections, we have noted a trend of publishing graphics with the alleged rating of politicians and parties, with the name of one of the well-known public opinion polling agencies attached. This method of disinformation is cheap yet effective. There are Internet websites where one types in names and data to get ready-made graphics that can be presented as results of public opinion polls and shared on Facebook, Twitter, and Instagram.
However, there are more subtle methods. The CSI methods. In the run-up to every election in Montenegro, a certain Centre of Independent Researchers (CSI) from Belgrade publishes public opinion surveys and projections of election results.
This time, CSI came out with a spectacular new methodology, so, as they claim, instead of the standard sample of 1,000 to 1,500 respondents, they conducted the survey on as many as 323,658 citizens of Montenegro. According to the Ministry of Internal Affairs’s official data of the, a total of 542,154 citizens have the voting right in Montenegro. So, the CSI called far more than half of the citizens with the right to vote in Montenegro by phone, or at least that is what they claim.
CSI stated that public opinion polls in Montenegro so far were conducted on the so-called representative sample of 1,000 people.
“How representative that sample is, can be seen from the fact that all those surveys turned out to be incorrect, i.e. they were the result of spins and manipulations. That is why there is no doubt that this survey, which was conducted on as many as 323,658 citizens, is accurate and relevant,” CSI claims.
Their poll was published by Borba, Adria, Alo, Kurir, Srbija Danas, Pogled, 24Inform.me, In4S…
According to that survey, Milo Djukanovic and Andrija Mandic will allegedly advance to the second round of the presidential elections. A common characteristic of all CSI surveys is that all of them project that the best results will be achieved by DPS and DF and their respective candidates.
Specialized in Montenegrin elections?
It is impossible to find any CSI surveys online except those related to the elections in Montenegro. Thus, in March 2011, CSI published the results of a survey related to the local elections in Berane through the Borba web portal, informing the public that the coalitions gathered around DPS and the Democratic Front have the greatest level of support. CSI was only partially right. DPS did win the largest number of votes in Berane, which brought them nine mandates, but the Socialist People’s Party won more mandates (7) than DF (6), which according to the CSI survey, had only 0.3 per cent less support than DPS.
The same CSI (not the Miami one) published a survey on October 16, just a week before the elections, which predicted that the DF would be the holder of power in the new local government in Podgorica.
According to the survey, 39.4% of citizens would vote for the coalition around DPS, which includes SDP, SD, LP, while DF would be supported by 19.7% of those polled. The third place would be reserved for the coalition around Democrats with 14.7%, while Europe Now would have the support of 12.4% of voters. CSI added that 8.8% of voters would vote for Prime Minister Abazovic’s Civic movement URA, while the SNP, the newly founded Preokret party and the Svetosavska party would remain below the census. The reality, however, denied CSI. The movement “Everyone for our City” led by the DPS won 38.1 per cent, the “Europe Now” movement won 21.7 per cent, and the coalition around the DF, “For the Future of Podgorica”, won 18.2 percent – which means that the DF is not the holder of power in the Capital City.
No website, no registration, no contact information
An internet search about the Centre of Independent Researchers yields no results other than their surveys. There is no website, no contact information, nothing… CSI is not even in the Serbian Business Registers Agency, which means that they are not registered as a company. They are not even in the Montenegrin Central Register of Business Entities.
After the publication of the alleged research, the editorial office of Raskrinkavanje contacted the editorial offices of both Adria and Borba with questions about whether they had any information about the Centre for Independent Researchers, whether they could direct us to the original research or the website on which it was published, for whom the research was done and who presented it.
We received a response that they received a press release and graphics from the email address centarsamostalnihistrazivaca@gmail.com. In the announcement that was delivered to those editorial offices, there was no information about who commissioned the poll.
The editorial office of Raskrinkavanje sent the following questions to the address from which the announcement was sent:
- Given that it is impossible to find anything about the Centre of independent ist developer (CSI) online and that the organization is not in the registers of the Business Registers Agency, can you tell us something more about the research team of the Centre of Independent Researchers?
- How long has the Centre been in existence? Who are the methodologists of your Centre?
- Given the fact that the Internet search leads to your surveys related to the elections in Montenegro only, can you refer us to your other surveys that are not related to the election processes in Montenegro?
- When was the pre-election survey conducted, and who commissioned it?
- Phone surveys – CATI, are described as appropriate for marketing research but not for high-sensitivity projects such as political opinion polls. Therefore, how was the sampling frame established?
- Which database did you get the numbers from?
- Does the sampling frame include all three mobile network operators as well as the landline operator?
- How did you ensure such a good response rate (323,658 citizens of Montenegro), that is, the response rate by phone?
- How was equal representation by socio-demographic characteristics ensured? By using quotas according to census results or some other way?
- Does the unweighted database have certain distribution skewness?
- Where else has that method been used and can you share a source that further describes it?
- Why did you decide on a sample of this size when a sample of 1,000-1,500 respondents is the global standard for the most significant academic research for countries the size of Montenegro and when the costs for the research method you mentioned are up to 200-300 times higher?
We have not received any answers.
An announcement for selected only?
Furthermore, several Montenegrin media, specifically the Vijesti web portal, MINA news agency and Pobjeda, confirmed to Raskrinkavanje that they did not receive the announcement published by the media outlets we mentioned. How CSI decided to whom they would send their survey data remains unknown, given that they do not send their announcements related to the Montenegrin elections to the Montenegrin media with the largest readership.
The director of the DAMAR agency, Vuk Cadjenovic, explained to Raskrinkavanje that he did not see the part of the research that refers to the methodology except for the sample size.
“There is no information about the time of data collection, nor about the margin of error. These are the basic elements that speak about the methodology and the reliability of the research itself,” Cadjenovic said.
He adds that it is unusual that the research was conducted on such a large sample as claimed, and adds that the standard in practice is proportionally distributed 1,000 respondents.
“It is not impossible to survey a sample of 300,000, but it requires significant resources and organization. There is also the question of whether carrying out research on such a large sample is profitable,” Cadjenovic claims.
Since we have not received answers from CSI, it remains questionable how an agency from Belgrade, which is not registered and does not have its own website, is paying to survey more than 300 thousand citizens of another country?
The media violated the law and the Code of Journalists
By publishing survey results 11 days before the elections, Adria, Borba, IN4S, Pogled and 24inform.me violated the Law on the Election of Councillors and Members of Parliament.
This Law stipulates that within 15 days before the election day, radio-broadcasting services and other media shall not be allowed to publish the results of public opinion polling, other surveys and analyses regarding the choices of citizens with reference to election results assessment.
Also, by publishing the public opinion polling without indicating when it was conducted and who commissioned it, the media violated the Code of Journalists of Montenegro, which stipulates that when publishing the results of public opinion polls conducted by polling agencies, the media should state the number of respondents, the date of conducting the survey, the identity of the person commissioning the survey, as well as the questions asked in it.