Is Russian sentiment turning against the war? More evidence from emojis
A couple of weeks ago, I asked whether Russians were turning against Putin by looking at how they, or at least Russian-speaking readers on some popular Telegram channels, were responding to stories about him.
I’m inclined to think that the evidence I presented, in the form of a number of negative reactions to stories involving Putin, isn’t overly strong. As far as I can see (although this is based just on casual observation), there isn’t an ongoing trend in the Telegram reactions that would suggest an anti-Putin turn. Maybe it was just an aberration, or maybe people were indeed annoyed that Putin was doing PR stunts while the army was retreating, but that annoyance was transient. Or maybe — definitely not to be underestimated — it was just trolling.
Of course, things have changed a lot in 19 days since I wrote that post. Russians are fleeing en masse to avoid the mobilisation that doesn’t seem so partial and the army continues to get beaten back. The mobilisation provoked anti-war protests, and a narrative has arisen according to which the average Russian person has had a change of heart, albeit not out of moral disgust at the war but from self-interest (personally, while I’m sure there are some Russians with exactly that attitude, I’m sure there are many others who’ve always been against it and one should be slow to condemn people living in an extremely brutal authoritarian state. But many would disagree, and perhaps they’d be right to.)
But this change in how things are going gives us a reason to check again whether we can notice a sentiment change. The devil here is really in the details: you need to find a suitable source of data that can be used to get reactions, but that’s not super easy. In the last post, I explicitly went through posts manually, but one can’t do that at scale, and it risk cherry-picking; and if one tries to automate it (say, by looking at reactions for any story containing ‘Putin’, one would have far too blunt an instrument (and Telegram would probably rate-limit you)). Happily, I found a potentially useful such source, and that’s what I’ll present today, showing that there is indeed some suggestive data for a recent change in sentiment.
TASS is a Russian news agency. Its Telegram channel has about 1/3 of a million subscribers (which, notably, isn’t that many — various propagandists have been half and 1.2 million). It started having emoji reactions to posts around March this year (I don’t know if that was its decision or if that was just when Telegram in general started allowing them).
Pretty much every day, it links to its rolling coverage. Moreover — and this is what enables one to automate the data collection — it does so using the same message:
24 февраля президент России Владимир Путин в ответ на обращение руководителей республик Донбасса принял решение о проведении специальной военной операции.
On February 24, Russian President Vladimir Putin, in response to an appeal from the leaders of the Donbass republics, decided to conduct a special military operation.
The text then gives a link to the rolling news page for that day.
Since around the 24th of June, these posts get a few hundred reactions per day. (It’s oddly discontinuous: before that date, none of them get more than 15 reactions; after that, eyeballing it, none get less than 200. I’m guessing the explanation is some update in Telegram but am not sure about this.)
So then here’s a natural question: how, over time, have the readers of this Telegram reacted to the rolling news? Well, here’s the answer.
Telegram allows several different types of reactions: thumbs up, thumbs down, smiley, fire emoji, worried emoji, and so on. The vast majority use thumbs up or down. There are, in addition, more positive than negative emojis. So, to compare like with like, I just looked at the thumbs up and thumbs down reactions for a given post.
With that said, here is first the graph showing date along the x and thumbs up along the y:
To my untrained eye, nothing particularly jumps out. The three outlying dates around the end of August are 23rd, 26th, and 3rd of September. A quick glance at both the Wikipedia time-line and the TASS posts for the day (here, here, here) didn’t reveal anything, but I’m pretty sure I’m missing something, as we’ll see if we look at the thumbs down reactions:
We’ll talk about the most obvious trend in a second, but note that that same end of August period also saw a spike in negative reacts. What was it? (Possibly at least part is Dugina’s death of 20 August, but that can’t account for all of it.)
The most interesting thing is definitely the spike that takes off around 20th September. And here I take it we do have an explanation: it was 21 September that Putin announced the partial mobilization. What is interesting is that the thumbs down seems to been reverting back to roughly their old level in the last few days.
Finally, it’s perhaps helpful to check out a single figure, arrived at simply by subtracting the negative from the positive reactions (so that a negative number means there were more thumbs down than up):
Arguably that makes more evident the relative constancy, minus some outlining and partially explanation days, in the reactions.
Lessons? I think probably the most reassuring (albeit not particularly sensational) is simply that the Telegram reactions seem to track real-world events, most recognizably in the reaction to mobilisation (more carefully: in the spike in negative reactions after the announcement of mobilisation).
And if then Telegram reactions are somewhat reliable, then that gives us arguably a good source of information about at least some Russian sentiment, and for example, we should take seriously the fact that the negative reactions seem to be trending down as extremely tentative evidence that the mobilisation didn’t permanently change sentiment. At least, I think the modest conclusion that this is potentially a decent source of data that should be paid attention to in future is warranted, and it’ll be interesting to see, after Putin’s planned annexation speech later today (happening, as it might, with another big defeat in the east.), what the reaction will be.