Why the Discussion of Putin’s Support Has Lost Its Meaning
An extended commentary by Grigory Yudin on a study that concludes that Russians’ actual support for Vladimir Putin is close to what sociological polls show
Доступно на русском
This article is an extended commentary by Grigory Yudin on a study that concludes that Russians’ actual support for Vladimir Putin is close to what sociological polls show. IStories previously told about this study, which other sociologists consider correct.
One of the main complaints about opinion polls on political topics is that dissatisfied people are afraid to answer honestly and refuse to participate in polls, which distorts the results (there may be many other reasons for this). To check this, the authors of the study conducted a “list experiment” — this method is often used to study people's reactions to sensitive topics. Respondents are divided into two groups. One group is given a list of three political figures, while the second group is given the same list plus Putin. People in both groups are asked how many politicians they support. Not who exactly, but just a number. Putin's support can be assessed by the difference in the responses of the two groups. For example, in January 2015, Putin's inclusion in the list added 0.81 (the average of responses in the first group was 1.11 and in the second group was 1.92), allowing us to estimate his level of support at 81%. This is about 5 percentage points less than what the opinion polls gave Putin at the time.
This is a strong simplification. There is an interesting effect: adding another figure to the list reduces the result (it is not who is added, but the very fact of lengthening the list). This “deflation” can be assessed by conducting another placebo-like experiment: there is nothing related to Russia. If we adjust the results of the list experiment for this artificial underestimation, we get numbers that are quite close to the polls. The authors of the study conclude from this (shared by many, but by no means all) that respondents do not lie to sociologists that much.
The study once again reproduces the thesis that Timothy Fry and his colleagues have been stating for a long time: that mass polls are absolutely correct about Putin’s high support in Russia. In one of the papers, the same team even managed to show that the polls underestimate this support — that is, Russians actually like Putin even more than the Pollsters tell us.
But even without regard to the numbers, there are several problems here.
What is “support”
The word is often used (including in Fry’s article) comma-separated with “popularity” and “approval,” as if it were all the same thing, and something quite obvious at that. However, the very assumption that everyone has a built-in sensor for supporting or not supporting Putin is completely at odds with the Russian political reality.
Russia is an extremely depoliticized country, and as soon as the conversation turns to politics, the interlocutor will certainly try to change the subject, feeling uncomfortable; just to avoid spoiling the relationship. Russians in general try to stay away from politics; there is a widespread belief in the country that nothing can be changed anyway, it is better not to have political views, and one should focus on what is at least partly in one’s power (family, career, consumption). When we ask such a respondent whether he “approves” of Putin’s activities, how does he understand this question?
Modern Russia is a plebiscite regime, where the question about the emperor is perceived not neutrally, but as a request for acclamation [an expression of approval of his actions]. In the United States, where some of the authors of the article work, the question of whether respondents approve of Donald Trump’s activities suggests that if they do not approve, they may, for example, approve (or disapprove) of Kamala Harris. There is nothing like that in Russia, and respondents are aware of it. That is, the question “Do you approve of the emperor?” is a request for a ritual demonstration of loyalty. This does not mean that the respondent is lying. It's like saying “Christ is risen!" in response to “Indeed He is risen!” — not because you suddenly realized at that moment, “He’s really dead, and the coffin is empty!”
The polls have completely convinced the political class both in Russia and abroad that Russians are in favor of Putin and love him. <...> This leads to a radically wrong assessment of the situation and corresponding decisions.
The meaning of “support” and the like is not obvious in these circumstances. Graham Robertson, for example, shows that this support is more than 80% correlated with what is known in psychology as “justifying the system” — with statements like, “The government is bad, but we don’t deserve any other,” “The government is bad, but any other government will be worse,” or “The government is bad, but the rest of us are no better off.” That is, when people approve, the meaning of their statement is roughly as follows: “I don't like it at all, but I don’t believe it could be otherwise.” Can this be called endorsement? Or popularity? Probably, if you define support as “disgust combined with acceptance in the absence of alternatives.” But why call black white?
Interestingly, the authors of the study under discussion are quite receptive to these considerations and recognize that "support" is a very conventional term used simply for convenience. The only trouble is that when these studies are released to the general public, the public is not informed about it, nor will they dig into such methodological subtleties. If it is said that Russians support — it means that they love the tsar!
This has a similar effect on politicians both in Russia and abroad. Surprisingly, the polls have completely convinced the political class both in Russia and abroad that Russians are in favor of Putin and love him. That is why every time the standard Russian reaction to political events - “ it’s none of my business” — comes as a surprise. Now it happened after the events in the Kursk region. “Why didn’t Russians stand up as one in defense of Putin? After all, we know from reliable research that they adore him, don’t we?” This leads to a radically wrong assessment of the situation and corresponding decisions.

Looking for something that is not there
All work with list experiments is based on the theory of “falsified preferences,” which was proposed by Timur Kuran. This theory proceeds from absolutely ridiculous assumptions about human political behavior. According to this theory, anyone has preferences on political issues, and can either state them honestly or lie in response to a question. Accordingly, list experiments attempt to calculate the proportion of those who have lied because they fear reprisals.
This picture of the world has nothing to do with the reality of the authoritarian countries for which it was invented. As discussed above, the default situation in authoritarian settings is not to lie, but to depoliticize. Why withhold one’s political views when it is easier not to have them?
Of course, direct lying by respondents out of fear also happens in Russian surveys. Experienced interviewers often assume that respondents are afraid, and it is doubtful that the interviewers are entirely wrong. Surely the situation must have deteriorated further since the war began.
However, fear is not at all the main problem in measuring support in Russia. People often have no preferences even on the most basic issues. Moreover, under authoritarian conditions, it is explicitly prescribed not to have any preferences - it is safe, does not make you a social outcast, and raises your self-esteem (“I am a serious person, I am not interested in politics”). Since the Russian regime is effective (and it is), a significant part of respondents have no preferences and formulate them right on the fly, oriented towards understanding what is expected of them. Since the poll is usually perceived as a situation of communication with generalized “power,” the question about Putin’s approval is read as a test of loyalty.
Observers are sometimes absolutely certain that all Russians must have an opinion on such key issues as “Putin's popularity” or “support for the ‘special military operation.’” But why do we need this opinion? How will disapproval or denial of support affect events? Will it make Putin cease to be emperor? Or will the war stop? No one believes in this, and rightly so, and therefore people prefer to believe that “everything is not so unambiguous.”
This is why all studies with list experiments are highly questionable — and not at all for the methodological reasons given in the article. All this literature has fictionalized rational and convinced voters who are deterred from expressing these beliefs only by fear. Of course, such people exist too — as these studies show, they are typically 5–10%. But people usually react to tyranny in a very different way.
An impossible task
In a depoliticized context, people prefer not only not to have opinions or express them, but also to shy away from interactions that are not directly relevant to their lives.
The Levada Center, which did the fieldwork for this article, used the personal interview method. This means that an interviewer comes to your home to ask you all these questions and conduct a list experiment with you. Imagine — some person is trying to get something out of you about politics and asking cognitively quite difficult questions (“I’m going to give you four judgments; tell me how many of them you agree with” is a very difficult question to understand). You may have been running somewhere between the bathroom and the kitchen, and now you have to count the number of statements you agree with.
The personal interview method is the riskiest because it leads to the most serious sampling bias. People who can be interviewed at home are very different from those who are not at home, who are shut out by a fence and a security guard, who do not open the door, who refuse to talk to the interviewer on the landing, or who refuse to let him or her into the apartment.
The validity of such an experimental design is rightly put into great doubt. In any case, convincing cognitive testing is required to show that it is at all solvable for the respondents.
***
Polls should not be used as plebiscites — the question “How many actually support the czar?” is meaningless. Polls, like any cognitive tool, can provide interesting and useful information. However, they exist in the political field, and it is naive to think that since they are conducted by scientists, they are the ones who report the whole truth, you just have to find the right figure and make sure that it is really it. That is why articles with headlines like “How many actually support Putin” will continue to appear. They do not help us understand what is going on in Russia. But they work perfectly well for those who have turned Russia into a continuous plebiscite.