Thread by @samagreene

Because the world clearly needs yet another 🧵 about whether Russians really support the war - and because I'm probably the only one left who hasn't commented on that @levada_ru poll showing 81% support - here's my take: Yes, but.

If you haven't seen the poll, the details are below. In a nutshell, @levada_ru's numbers show 81% support, highest among those aged 55+ (86%), lowest among those aged 18-24 (71%).
levada.ru/2022/03/31/kon…

To be honest - and I've hesitated to write about this, out of respect for my friends at @levada_ru - I would not have published this survey. Not because the results are unpleasant (though they are), but because they are impossible to interpret.

Broadly, my objections to this survey are the same as they have been to surveys throughout this war, as I summarized in a thread on 5 March:

https://twitter.com/samagreene/status/1500163777878188034

Mostly, these problems fall into two categories:
(1) Response validity; and
(2) Representativeness.

Response validity is always a problem in surveys, but it's particularly a problem now. At its core, the issue is that people have choices when deciding how to answer a question, and those choices only partially reflect their actual opinions.

Responses can also reflect social pressure, fear of consequences, etc. All of those things should be expected to affect responses right now, and there are no devices in the survey that would allow us to control for them.

In other words, we know that 81% of respondents said they support the war. But we don't know _why_ they said that, and we can't know it.

On representativeness, other survey researchers have reported response rates of ~5%. Depending on the survey, that's about 1/2 to 1/3 of the response rate we'd usually expect. Alas, @levada_ru haven't published their response rate.

There is every reason to believe that there is a direct relationship between opinions about the war and willingness to respond to the survey, so publishing the response rate for a survey like this is vital.

And unless @levada_ru can demonstrate that this survey's response rate is broadly identical to pre-war response rates across its stratification variables -- age, gender, geography and education -- there is simply no way to argue that the results are representative.

All of that said, let me dismiss two criticisms I've seen leveled at this survey:
(1) It was not vetted, approved or censored by the Kremlin;
(2) The numbers were not doctored.

So does all of the above mean that the survey is meaningless?

No, it doesn't.

While methodological problems mean we cannot say anything meaningful about the actual breadth of support, it seems clear that the broad trend is real: Russians _are_ rallying around the flag.

Take a look at two other indicators @levada_ru tracks. First, respondents' sense of whether the country is going in the right direction (dark blue) or the wrong direction (light blue):

And Vladimir Putin's job approval ratings (dark blue = approve; light blue = disapprove):

Both of these indicators have returned to levels last seen during the long rally around the flag sparked by the annexation of Crimea in 2014.

To be clear, I am skeptical of the point values here, but not of the general trend. The Kremlin _has_ been able to manufacture a significant shift in public opinion.

There are other indicators of this, too. Putin's 18 March rally at Luzhniki reflects both the Kremlin's need to drum up support, and its ability to do so -- an ability it did not have in the early days of the war.

This is in part because we are seeing much more concerted coverage on state television, something that was absent in the first few days of the invasion.

We're also seeing schools, churches and other quasi-social institutions swing into action, creating rituals of participation -- such collecting letters and care packages for the front.

And while "rituals of participation" might sound like a softy sociological kind of thing, they are really the crux of a rally around the flag.

As @gbrunc and I showed in our research on the post-Crimea rally, the massive upswing in public support for the regime is predominantly an emotional affair.

cambridge.org/core/journals/…

The emotions that underpin a rally are generated not by a vertical connection to the leader, but by horizontal connections to fellow citizens. That's why these "rituals of participation" matter -- they embed people in a political community infused with emotion.

That said, this rally is different. In the early days of the war, @russophiliac aptly described what was emerging as "defensive consolidation", rather than a rally, and I think that was right.
postsocialism.org/2022/03/04/def…

While I do think we're now seeing a genuine rally, the "defensive" aspect Jeremy noted is key, and it's an aspect that wasn't present in 2014. As Graeme and I showed, that rally was all about pride and hope. This one is about anger and fear.

Also unlike the Crimea rally -- which was mostly a spontaneous, bottom-up process of political consolidation (though supported by propaganda) -- the 2022 rally is a top-down process, instigated by a combination of propaganda and repression.

I'm going to have to spend some more time thinking about the difference between prideful/hopeful rallies and angry/fearful rallies, but my initial impression is that anger and fear will draw sharper lines between "us" and "them". That will add to Putin's resilience.

On the other hand, rituals focusing on negative emotions may be harder to sustain than rituals focusing on positive emotions. It thus seems unlikely to me that this rally will last anything like as long as the four years of the Crimea rally.

To sum up, then, I think we have to accept that while the polls are problematic, they are reflecting -- if imperfectly -- a social reality, and that social reality is a rally around the Russian tricolor.

There is reason to believe that this rally will not last very long, and to question its social breadth and depth. But based on the evidence available, I see no reason to question the reality of the rally itself. Alas.

/END