In 2022, tensions increased around opinion polls in Russia, with many scholars publicly denying their reliability. This paper explores the ongoing reassessment of survey research in Russia by showing how polls on war support have evolved, as well as how they are portrayed by different actors in and outside academia. Using a dataset of news media publications, I inquire into ways in which surveys are portrayed as either trustworthy or not, depending on their results and the source they come from. I demonstrate the tension between scientific and political discourses – e.g. if a poll shows the majority support the war, its opposers blame methodology for the results, and vice versa. I claim that the observed methodological issues were always present in the Russian survey research and it is the transformed sociopolitical context that has made them more visible and frequently discussed. The methodological discussion about the reliability of poll figures intersects with the political discussion about how many Russians support the war. The latter unfolds not only as an empirical but also as a political question. Examining opinion surveys in the context of the extreme event of the Russian war against Ukraine will help develop a revised approach to survey measurement on sensitive topics in politically and socially constrained environments, and show how to make better sense of the limited data at hand, as well as inquire into ways in which polls are portrayed in media as either trustworthy or not.