This week, Russian news reported that trust in President Vladimir Putin fell to its lowest level in 13 years. Or did it? A monthly poll collected by the Kremlin-friendly polling organization WCIOM has long asked Russians, “Among politicians, who would you trust and who would you not trust to make important government decisions?” In May, just 31 percent listed Putin as a politician they would trust — the lowest figure since 2006.

Dismayed with the results, the Kremlin asked pollsters to investigate the causes. In response, WCIOM commissioned a new survey and asked the direct question: “Do you trust or not trust Vladimir Putin?” Seventy-two percent of respondents answered yes.

How should we understand these two results? Obviously, we should be concerned about a pliable public opinion firm bending to the Kremlin’s will, with political motives in changing the question’s wording and with public confusion about the results. The Kremlin is keenly interested in public opinion toward Putin. And given the variation in responses to these questions, we must wonder about the usefulness of public opinion polling in Russia.

Declining trust

Polling organizations know that different formulations of similar questions often generate different answers. In fact, small changes in question wording and response options can have large effects. What makes the first question valuable is that it has been asked repeatedly over time, allowing us to establish a trend. When asked the same question in May 2018, 48 percent of respondents listed Putin as a politician they would trust. The drop from that to 31 percent is large. We usually gain more by comparing responses to the same question over time than by trying to interpret answers to slightly different questions.

Second, both forms of the trust questions can provide useful information, but they ask a respondent to do different things. Asking a more direct question will generally lead to higher responses than asking respondents to list items off the top of their heads. Direct questions force respondents to answer. List questions without any prompts allow a respondent to avoid answering the question directly. WCIOM appears to have chosen the direct question about trust in Putin in hopes of showing higher levels of trust in the president.

The direct formulation that asks whether you trust Putin is similar to the presidential approval question commonly asked in Russia: “On the whole do you approve or not approve of the activities of the President of Russia?” There is good evidence, albeit a bit dated, that Russians are telling the truth when they are asked whether they approve of Putin’s activities as president. WCIOM reports that Putin’s presidential approval ratings in May were 65 percent, much like the 72 percent who, when asked directly about whether they trust Putin, said yes. And it’s similar to the more independent Levada Center report that 66 percent of Russians approved of Putin’s performance as president in May. The decline in trust in Putin measured by WCIOM in the list question also mirrors a drop in presidential approval in the past 12 months captured by both survey organizations.

Third, digging deeper into the data reveals a trade-off between the two forms of the trust question for the Kremlin. WCIOM’s list question reveals that just under a third of respondents listed Putin as a politician they would trust, but only 7 percent listed Putin as a politician that they do not trust. The direct trust question finds that 72 percent of Russians trust Putin, but also that 24 percent of Russians do not trust Putin. The direct trust question reveals not only more trust in Putin, but also more distrust.

Trust no one

Finally, perhaps most dismaying for the authorities, is how little respondents trust any politicians in Russia. When asked to list politicians that they trust, the second- and third-place finishers were Defense Minister Sergei Shoigu — with 15 percent — and Minister of Foreign Affairs Sergei Lavrov — with 13 percent. Just 8 percent listed Prime Minister Dmitry Medvedev as a politician they trust. And in the direct question, 58 percent of respondents said that they do not trust Medvedev. Respondents may be more comfortable admitting distrust of Medvedev than Putin, but the overall picture for the Kremlin is not pretty.

Interpret all this carefully

Public opinion polling in Russia (and elsewhere) is beset with problems, but it can tell us a lot about politics when we interpret the results with care. And given WCIOM’s willingness to bow to Kremlin pressure, we will need to take even more care going forward. Indeed, the larger story to watch may be the threat to the quality of public opinion polling in Russia. This would be a big loss. Public opinion polling has been far better in Russia than in most autocracies. Ironically, it would be a loss for the Kremlin too — as autocrats also need accurate information about the public mood.

But it seems reasonable to conclude from the data that trust in Putin is declining, albeit from a high level, and that trust in other politicians is very low. When thinking about the prospects for political change in Russia, this is valuable to know.

Timothy Frye is the Marshall Shulman professor in the department of political science at Columbia University and is finishing “Weak Strongman: Why We Get Russia Wrong and How to Get It Right.”