On February 3, 2019, Channel 24 published a piece titled “Why the ratings of presidential candidates differ significantly: an explanation”. Instead of an analysis of trends, the piece demonstrates the art of manipulation by certain journalists using sociological data. Given the content and tone of the story, one may assume that its purpose is to discredit sociology in general and specific sociological organizations in particular. We also sincerely hope that this story does not reflect the editorial policy of Channel 24 and represents only the opinion of individual journalists. Why do we believe this?

Fact one.
The article accompanying the story states that “the Director of the Democratic Initiatives Foundation claims that so many different surveys are published so that people do not see an objective picture.” In reality, these are not her words — she was reading out a message (a post) written by a journalist. The authors of the story repeatedly use quotes by Iryna Bekeshkina and Oleksii Antypovych from the round table “Sociology and Elections: an argument for a rational choice or manipulation?”, taking them out of one context and placing them into a completely different one.
Conclusion: manipulation.

‍

Fact two.

A quote from the story states that, having compared the ratings of two sociological groups, journalists noted that during the 2014 presidential election Petro Poroshenko was given quite different percentages before the first round. According to the Democratic Initiatives Foundation and the Razumkov Centre, he was given around 45%, while the Rating Group gave 34%. The authors claim that a ten-percent difference is substantial and exceeds statistical error, and further compare these figures with the actual election result, where Poroshenko received nearly 56%.

Conclusion: manipulation. The authors compare indicators that are not comparable and on this basis construct manipulative conclusions.

First, they compare the Democratic Initiatives figure of 45% among those who intended to vote with the Rating Group figure of 34% among all respondents, concluding that the difference is significant. In reality, among all respondents the figures were 35% and 34% respectively — meaning that the difference is фактично absent.

Second, they compare these figures with the actual election result. However, in the 2014 presidential election there were 29.6 million registered voters, about 18 million participated, and 9.9 million voted for Petro Poroshenko. This equals 54.7% of those who voted, but only 33.4% of all registered voters. Thus, the Rating Group not only did not miscalculate its forecast, but predicted the result quite accurately: 34% among all respondents versus the actual 33.4%.

For correct comparison between survey results and election outcomes, the Rating Group specifically calculated and published the indicator “percentage among those who intend to vote and have decided on their choice.” This indicator was available in May 2014, yet the journalists did not refer to it at all.

We could assume that the authors simply confused different indicators (among all respondents, among those who intend to vote, among those who have decided). However, similar theses had already appeared on social media, promoted by certain commentators, such as claims that “your May 2014 rating gave Poroshenko 34%, but he received 55%,” or that “sociological surveys convincingly prove that the Ukrainian people are twice as smart as sociologists think.”

‍

Fact three.
Another quote from the story claims that different sociological organizations give different outcomes for a hypothetical second round between Petro Poroshenko and Yurii Boiko, and presents a graphic titled “Ratings of sociological centers regarding the second round of the 2019 presidential election.”

Conclusion: manipulation. The authors compare data obtained at different times and construct conclusions on this basis. For example, the most recent data published by the Rating Group in January 2019 show that in a second round Poroshenko would receive 23% and Boiko 21% among all respondents. Meanwhile, the Kyiv International Institute of Sociology had not published any second-round data in 2019 at all.

We therefore consider such actions by the journalists — namely, the authors of this Channel 24 story — to be unprofessional from a journalistic standpoint, but quite professional from the standpoint of manipulation.

On February 4, Channel 24’s website published another analytical piece, “How politicians’ ratings in Ukraine changed after the elections: infographic”, which was based exclusively on data from the Rating Group. We have never prohibited journalists from using our data; on the contrary, we welcome genuine analysis. This article itself demonstrates that Channel 24 journalists understand that election results can only be compared with candidate ratings among respondents who have decided on their choice. This is exactly how the infographic was constructed. However, in the earlier story “Why the ratings of presidential candidates differ significantly: an explanation”, the authors compared election results with ratings among all respondents.

Even in this infographic, however, a number of errors were made. For example, figures for November 2016 are given among all respondents (and even then correctly only for Tymoshenko and Poroshenko), while figures for other months are given among those who have decided, and then presented as a single “dynamic.” In such cases, we recommend using a single primary source — the website of the Rating Group — where candidate dynamics are calculated using a unified methodology, ensuring correct comparison. There is also always the option to contact the Rating Group directly for clarification from the primary source.

Therefore, the Rating Group demands that Channel 24 remove the mentioned materials and issue an official retraction of the manipulations disseminated in the story “Why the ratings of presidential candidates differ significantly: an explanation.”

Otherwise, Rating Group will cease any cooperation with Channel 24.