So…SoupGate is still a thing right? I wanted to clear up a few things since my first post about it, and include some counterarguments.

First of all, I think it’s important to clarify what we’re trying to actually achieve with these actions. There are many potential outcomes that protest movements seek, whether that’s policy change, growing the movement, shifting public opinion, or influencing public discourse. If you’re interested in how protests can achieve these various outcomes, check out our protest outcomes report.

In this case, a lot of the debate has been around whether public support for climate action, or the climate movement, has been impacted positively or negatively by SoupGate. I’ll focus on public opinion in this piece, as I do think it is one of the most important ways for protest movements to achieve long-lasting change. That said, the main impact of these radical actions might be creating more “hardcore” activists, which could be a success metric in its own right.

The key debate going on in the Twitterverse is whether these actions positively or negatively impact public opinion on climate action. For people who think it’s had a negative impact (e.g. Michael Mann), they’ve been sharing an interesting paper by Feinberg, Willer & Kovacheff (2020) which is appropriately titled “The Activist’s Dilemma: Extreme Protest Actions Reduce Popular Support for Social Movements”. The claim here is that extreme actions, like throwing soup at a Van Gogh painting, actually reduce support for the movement (as well as identification with the movement).

First of all, I think this is an interesting paper, and I’m really glad the authors did this research. There is such little empirical evidence around effective tactics, and this one is a total gem that compiles over 6 different experiments in one paper! That said, I think people are quoting some of the findings without fully considering how the study could apply to the real world.

To start with, the impact studied in this experiment is only part of the full story. By ensuring that everyone taking part in the study is exposed to either the moderate or radical treatment, we’re essentially controlling for issue salience (e.g. the number of people who hear about an issue is the same in both cases). Whilst this experimental approach is extremely useful for determining causality, it isn’t how it works in the real world. One would expect the radical tactics to get many more views relative to moderate tactics, and this is a very important factor in the overall impact of an action or campaign.

To simplify things, we can model the impact of a protest on public opinion as: Impact on public support for the climate movement per view multiplied by the number of views. Here when I say public support, I’m talking pretty broadly and include things such as willingness to take part in any form of climate action, support for climate policies, support for the moderate flank, etc. For example, we can compare what the total impact might be in the two diagrams below, when considering the different effects of moderate vs radical protests (in an idealised best-case for radical protest).

So even though the impact on public support per view is lower in the radical action frame, the number of views (or amount of media coverage) is generally much greater, such that the overall impact is larger. This is something that sadly isn’t able to be captured by experimental approaches, and is best suited for public opinion polling approaches.

Additionally, this paper doesn’t use a control group with no protests happening for 5 out of 6 experiments. Robb Willer (one of the authors) and I had a pretty extensive DM conversation about this, and I think there are good reasons why the paper was designed in this way. As he points out, social movement groups aren’t usually thinking “Should I protest or should I do nothing?” but rather “Should I protest in a moderate way or in an extreme way?”. So in that sense, a control group might not be particularly informative.

However, without the existence of a control group, it’s impossible to determine whether doing an extreme protest is worse than doing nothing at all, which (in my opinion) was the claim being made by many commentators. Most were saying that these actions are overall negative for the climate movement, rather than just being worse than another alternative (e.g. moderate protest). For example, even though the impact of extreme protest is less relative to a moderate protest, it might still be positive. In fact, that is exactly what Bugden (2020) finds, which you can also see in the chart below.

As you can see, although the disruptive protest (civil disobedience) was less effective in raising support relative to the peaceful protest, it was still better than nothing. Of course, there are plenty of limitations to extrapolating from this one study, as it literally is just one study. There are some other studies, like Shuman et al. (2020) or Study 3 within the same Feinberg et al. paper that look at no protest vs disruptive vs violent action (with the first one favouring disruptive protest and the second one not). The findings from Study 3 from Feinberg et al. can be seen below.

As highlighted above, the authors do find that extreme protests are actually detrimental to the aims of the protests (e.g. counter-Trump protests actually increase support for Trump).

However, a key thing to note here is how the authors construct an “extreme” protest in this case, which is:

“Participants in the extreme protest condition watched news coverage of anti-Trump protesters gathering in the middle of a busy street, physically blocking carloads of Trump supporters from reaching a Trump campaign event and causing a traffic jam. The reporter covering the event describes the protesters as creating “a potentially dangerous situation” because their “actions are causing motorists to drive into oncoming traffic.” [emphasis mine]

In my opinion, this feels like it's straying into the more violent territory, as innocent human lives are now at risk due to this protest. However, it is still technically nonviolent and it does involve primarily road-blocking, which is a very commonly used tactic with disruptive protest movements.

The most important distinction might be: How do we generalise from experiments that show people blocking roads to throwing tomato soup at a painting? Or to literally any other forms of protest that we haven’t yet tested experimentally? This is where we have to use our intuition and subjective judgement, and where people can very reasonably disagree.

Essentially, based on evidence and good reason, we have to make a judgement call on whether we think the impact of a specific radical, disruptive or extreme action is overall negative or positive. The sign of the average individual exposure is very important, as it significantly determines whether people think of the action as very good or very bad, as seen below. I also think we need to accept that some actions can be too “radical”. For example, despite these actions being nonviolent, I think most would agree that XR’s Canning Town action was likely damaging for the climate movement, and I’m sure most also think that pouring human faeces on the memorial of a loved UK veteran is not that great either.

So what do others think the overall impact of this particular protest was? Some knowledgeable people I respect a lot think it was negative (e.g. Robb Willer & Winnifred Louis) and others more positively (Dana Fisher). I encourage people to check out their respective thoughts, particularly the very extensive thoughts from Dr Louis here.

A few key arguments that could make this action negative overall is that:

The radical flank effect only works when the public can distinguish between the radical flank and the moderate flank. I think this largely isn’t an issue, as the climate movement is extremely big and well-known, with popular moderate groups like Friends of the Earth or WWF.
This action crossed some threshold of being “too extreme”, similar to dumping faeces on a memorial, such that the impact was negative. I’m sympathetic to this argument, and it definitely could be the case, but I also think it’s almost impossible to prove. For example, it depends where the protest would lie on this graph I made up below (which is very rough and obviously context dependent but I think the underlying message is in the right direction)

Overall, I’m quite unsure if it was overall good or bad for the climate movement. I find some of Dr Louis’ points in her Twitter thread quite compelling, as well as others like Robb Willer pointing out the minimal “action logic” or narrative behind this action. As is a common trope in academia, more research is needed!

Understanding Social Change

Why I could be wrong about SoupGate