Thanks for this work and the nuanced write up.
I guess one way to make these findings potentially stronger / more informative would be to see whether effect size varies as one would expect with familiarity of Just Stop Oil (i.e. more change with more exposure)?
Thanks Johannes!
We thought about doing this but ruled it out as there would be a pretty clear bias e.g. the people who are most likely to hear about Just Stop Oil are people who are climate-conscious already, and are therefore more susceptible to positive shifts. I think we did do this informally, and did find a positive correlation between knowledge of Just Stop Oil and the constructs, but I don't think it's particularly robust.
Thinking out loud but maybe one way to control for this might be doing this within groups of people who answered the same to "How concerned are you about climate change" in the first survey, although this might make our sample sizes quite small / no longer representative.
Note: This is our more accessible report, with a more technical one intended for academic publication, with additional analysis, being released soon. Subscribe to our newsletter if you want to read the technical write-up at a later date. If you want to skim this analysis, we recommend you read the summary, key results and discussion.
Social Change Lab commissioned YouGov to conduct three nationally representative surveys of approximately 2,000 people each in the United Kingdom, focusing on their views on climate change, their willingness to engage in environmental activism, and their opinion of the climate activist group Just Stop Oil. The surveys were conducted before any significant protest activity occurred (on March 29th), during the Just Stop Oil and Extinction Rebellion protests (on the 9th of April), and after most (but not all) of the protest activity had occurred (on the 19th of April). Despite disruptive protests, there was no loss of support for climate policies, providing some evidence against the notion that disruptive protests tend to cause a negative public reaction. Furthermore, our findings suggest that the protests had a marginally statistically significant effect (p = 0.09) on respondents’ self-perceived likelihood of participating in environmental activism. Our data suggests that the number of people saying that they were willing to take part in some form of climate activism increased from 8.7% to 11.3% of the UK population, equivalent to approximately 1.7 million additional people. However, due to some unforeseen circumstances, we’re unsure on exactly how to interpret the strength of these findings on willingness to take climate action (explained below).
Whether protest can have an impact on public opinion is an important question for figuring out the impact of protest more broadly, as the impact of protests on legislators and its ability to influence policy is likely to be at least partially mediated by the impact of protest on public opinion (Agnone, 2007). A meta-analysis by Burstein (2003) examines the relationship between public opinion and policy change. He finds that in 75% of cases of policy change, public opinion plays a significant role. With this in mind, it seems especially important to study the impact of protests on public opinion, given the impact of public opinion and policy and legislator behaviour.
Whilst there has been some analysis of the impact of protest on public perceptions for a certain issue in experimental conditions (Bugden 2020, Feinberg et al. 2020), there has been little research examining the impact of large-scale protests using observational techniques, such as public opinion polling (Kenward & Brick, forthcoming). This seems particularly important, as the exposure to protest in experimental conditions is quite distinct to what is observed in real life. Experiments generally involve a single exposure to a neutral news item about protest, whereas reality may involve people being exposed to partisan media sources, word of mouth, and other sources of information. There has rightly been a criticism of experimental designs on the basis of the difference between experiments and reality. Observational work around ongoing protests provides a way for us to examine and more accurately evaluate the impact of protest on public opinion.
This was our goal for this research - using representative polling to better understand how public opinion around climate change and fossil fuels was changing as a result of protest and nonviolent direct action by groups such as Just Stop Oil and Extinction Rebellion UK. For further information on the actions of Just Stop Oil and Extinction Rebellion, you can read this article on Just Stop Oil or these highlights from Extinction Rebellion. Generally, the protests involved illegal actions such as sustained and high-profile oil infrastructure blockades (e.g. refineries and oil depots), glueing to government buildings, and blocking buildings of carbon-intensive industries.
Each of the three surveys had approximately 2000 respondents - the polling conducted before the protests had 2057 respondents, the polling during the protests had 2030 respondents, and the polling that took place after most of the protests had occurred had 2160 respondents. Further detail about our methodology can be seen in the Appendix or in our pre-registration. It should be noted that other environmental protests also occurred in April, so any change in views should be understood as the impact of the combination of activity from both Just Stop Oil and Extinction Rebellion, as well as other environmental protests.
Responses were weighted by demographic information (see further details in Appendix) in order to ensure that the survey was nationally representative.[1] All of the data can be found here. For context, we worked closely with Dr Ben Kenward, Senior Lecturer in Psychology at Oxford Brookes University, who has done similar research, as well as polling experts YouGov in designing our methodology and survey. We also sought feedback from those we considered survey experts, or had experience in survey design and public polling.
Respondents were asked “How likely or unlikely are you personally to do each of the following within the next 12 months?” and were able to rate themselves as being anywhere from ‘Very unlikely’ (1) to ‘Very likely’ (7). The types of activism they were asked about were:
Figure 1: Average Likelihood of taking climate action in the next 12 months, grouped by the period in which the respondent was surveyed.
For this aggregated metric, there was a marginally statistically significant change (p = 0.09) after adjusting for multiple hypothesis testing. Also of note, the change was statistically significant (p=0.02) before the adjustment was made.[2] Despite being outside standard scientific margins of error (p < 0.05), we choose to report this finding as in an applied setting, weak evidence can be more useful than not presenting any evidence, and that false negatives can be as damaging as false positives. The detected effect size was 0.07 (95% confidence interval (CI) 0.01 to 0.14), although the p-value was 0.09 after adjustments were made.
Additionally, we believe that there is a significant chance that the study had a sample size too small to detect a clearly significant effect given the effect size that we detected.[3] Given the small effect size that we found, a larger sample size would have been necessary to reliably detect an effect. We believe that the effect size in this instance was smaller than it might otherwise have been if not for the Russian invasion of Ukraine dominating the media (see Discussion).
Overall, we believe that this should cause the reader to weakly update in the direction of the protests having resulted in respondents reporting a higher likelihood of engaging in some forms of environmental activism. However, we are aware of the fact that the change is also plausibly due to noise, given the number of hypotheses being tested.
Figure 2: Average Concern About Climate Change, grouped by the period in which the respondent was surveyed.
For the second aggregated metric, which looked at one question about whether respondents believed that ‘climate change is one of the greatest threats to humanity’ and another that asked whether respondents felt ‘frightened at the prospect of climate change and the impact it might have’, there was no statistically significant difference in means. This indicates that the protests probably failed to make any impact on the public’s views about climate change. The detected effect size was -0.01 (95% CI -0.07 to 0.06).
Figure 3: Average Support for Just Stop Oil’s goals, grouped by the period in which the respondent was surveyed.
A third aggregated metric looked at the number of people saying that they supported the goals of Just Stop Oil (note that the questions did not specifically allude to the fact that these were the goals of Just Stop Oil, as we believe it’s more important to understand public support for the issue rather than the organisation). These questions asked people whether they supported or opposed three measures that the government could take: stopping the exploration of oil and other new fossil fuels, no longer granting new licences to oil and gas production, and making companies who extract fossil fuels spend their profits on transitioning to a new low-carbon economy. There was no statistically significant change in the mean response to the questions from before the protests to after the protests - the mean increased from 4.61 to 4.67, within our margin of error. The detected effect size was 0.04 (95% CI -0.03 to 0.10).
Figure 4: Awareness of the impact of fossil fuels on climate change, grouped by the period in which the respondent was surveyed.
The surveys also looked at the number of people saying that they were aware of the impact of fossil fuels. Specifically, respondents were asked whether they agreed or disagreed that ‘Fossil fuels are major contributors to climate change and environmental damage’. There was no statistically significant change in the mean response to the questions from before the protests to after – the mean increased from 5.13 to 5.16. The detected effect size was 0.02 (95% CI -0.04 to 0.08).
Figure 5: Average opposition to the goals of Just Stop Oil, grouped by the period in which the respondent was surveyed.
Respondents were asked questions that were intended to provide insight into whether the opposition to Just Stop Oil’s goals increased. One concern about disruptive protests is that they may be liable to increase opposition to the aims of the protesters. For instance, polling after Insulate Britain protests showed that many people in the UK believed that their traffic obstruction was likely to lead to more people opposing the aims of the protesters.[4] Hence, our surveys asked questions intended to gauge whether this is true in the case of Just Stop Oil’s protest activity. Respondents were asked whether they agreed or disagreed with two statements. The first statement was “Given the current high fuel prices, it's not appropriate for the UK to reduce extraction of oil and gas”, and the second statement was “Because of the war in Ukraine, we need to maintain our country's own production of oil and gas”, and responded on a 1 to 7 scale (1 being ‘Strongly Disagree’ and 7 being ‘Strongly Agree’). There was no statistically significant change in the number of people who were opposed to Just Stop Oil’s goals from before the protests to after the protests. The detected effect size was -0.01 (95% CI -0.07 to 0.05).
Figure 7: Support for Just Stop Oil in the survey that was conducted on the 19th of April, after most of the protests had occurred.
Figure 7: How much respondents had heard about Just Stop Oil in both the second survey and the final survey. We didn’t ask whether anyone had heard of Just Stop Oil in the first survey, as it was likely close to 0 due to few public actions to date at that time.
Evidence from these surveys suggests that most people do not support the actions of Just Stop Oil, even if they support the aims of the protesters. 56% of respondents in the survey that took place after the protests had occurred said that they opposed the Just Stop Oil protests, whereas only 18.1% of respondents said that they supported the protests. The likelihood of supporting the protests was correlated with which party the respondent had voted for - only 4.8% of people who voted Conservative in the 2019 General Election said that they supported the protesters, whereas 33.7% of Labour voters supported them.
Figure 8: Aggregated metric on likelihood of taking climate action, grouped by age group.
Unsurprisingly, younger people were disproportionately likely to say that they were willing to engage in some form of environmental activism in the next 12 months (See Figure 1). In the April 19th Survey, the mean response on the 7-point scale was 3.54 for Under 25s, whereas for Over 70s it was 2.61.
Figure 9: Changes in mean ‘Likelihood’ Score by Party from the survey before the protests to the survey that took place after most of the protests had occurred.
People who voted for different parties in the most recent UK General Election expressed differing views on how likely they were to engage in action related to climate change in the next twelve months, as can be seen in Figure 9.
No detected backfire due to disruption
The data presented suggests that it is unlikely that there was a strong backfire effect of the protests. There was no evidence that protests reduced support for climate policies or the goals of Just Stop Oil. Despite the fact that most people were opposed to the protests themselves (only 18.1% of respondents said that they supported the protests), there was no significant negative change in the number of people saying that they supported the goals of Just Stop Oil, or any fall on any of the other metrics that people were surveyed on. This seems relevant, as some experimental findings (Feinberg et al. 2019, Polanco et al., 2022) show a loss in support for an issue when protests are perceived to be “extreme”. Further research seems important to test how the public responds to different levels of disruption. However, as the protests by Just Stop Oil were particularly disruptive and disliked by the public, our findings indicate a group can be quite nonviolently disruptive without necessarily alienating people from your issue.
Increased willingness to take climate action
Our findings seem to suggest that the protests likely led to a small increase in the number of people saying that they were likely to engage in environmental activism in the next twelve months. This finding corroborates previous research by Kenward & Brick (forthcoming) that found that climate protests do have a significant impact on public opinion. Climate campaigns that use disruptive protest and garner large amounts of media attention seem to have a detectably significant impact on UK public opinion for climate change. It’s important to note that our findings have smaller effect sizes relative to Kenward & Brick (forthcoming), who also found statistically significant changes in constructs that we did not – such as concern for climate change and support for disruptive methods used by Extinction Rebellion. Other polling also indicates changes in public opinion due to Extinction Rebellion being much larger in 2019 relative to what we see now. One potential explanation for this is a “ceiling effect”, explored further below.
That being said, it should also be noted that the change in peoples’ self-perceived likelihood of engaging in environmental activism was only marginally significant (p = 0.09). We believe that our study may have had an insufficient sample size on account of the effect size being lower than we had anticipated. This may have been due to these protests receiving relatively less attention in comparison to previous high-profile protests in the UK, such as Extinction Rebellion’s April Rebellion in 2019. We explore two possible reasons for a small effect size below, namely lack of media coverage and the ceiling effect.
Lack of Coverage
Figure 10: Comparison of Google Searches in the United Kingdom for “Extinction Rebellion” (Blue) and “Just Stop Oil” (Red)
The protests by Just Stop Oil and Extinction Rebellion received significantly less media coverage than similar climate protests that have occurred in the past. It may be the case that these protests had a significantly lower effect size than anticipated because the media attention that they received was much lower than was the case with previous high-profile environmental activism. The protests received significantly less attention than previous climate protests in the UK, such as the Extinction Rebellion’s April 2019 Rebellion and Insulate Britain’s traffic obstruction that took place in September 2021. Figure 10 above shows the Google Trends data for searches for both “Extinction Rebellion” and “Just Stop Oil” - at the point at which Just Stop Oil was at its most searched, there were only 3% as many Google searches as the number that Extinction Rebellion received when it was at its most searched. Prior to conducting this study, we expected that the Just Stop Oil protests would receive more attention than they did, due to the significant degree of attention received by previous environmental activism. One likely explanation of reduced media coverage during our survey period is the near nonstop coverage of the Russian invasion of Ukraine.
Ceiling Effect
Additionally, it's plausible that UK public opinion on climate issues is already so high for certain variables, marginal increases are quite hard to elicit, which we’ll refer to as a “ceiling effect”. This is discussed briefly in Kenward & Brick (forthcoming), where “concern for climate change” had a mean response of 5.3 out of 7 initially, meaning there is little scope for this to increase. As the salience of climate issues has increased dramatically since 2018, it’s plausible that there are now diminishing returns on climate campaigns where the main outcome is high levels of media coverage, as most people interested in climate issues in the UK have likely already been exposed to relevant news and information. It’s important to note that a large increase in concern for climate change may have occurred due to the work of Extinction Rebellion, Fridays for Future and other grassroots groups, as explored both here by YouGov and here in our previous work.
To illustrate this ceiling effect, in our ‘Before’ polling we found that the mean score of “concern for climate change” was 4.8 out of 7 and the public awareness of the impacts of fossil fuels on climate change was 5.1 out of 7, implying high existing knowledge of the issues that Just Stop Oil was trying to draw attention to. A caveat is that we believe this is potentially only relevant to the UK and a select few other countries, specifically those with the highest percentages of public concern for climate change globally. One implication of this is to prioritise (potentially disruptive) climate protest and campaigning in countries with relatively low levels of concern for climate change, or, if protesting in countries who already have high concern, focusing on specific issues that have less public salience.
The activist’s dilemma: Preference for non-disruptive vs disruptive protest
It should be noted that there were three forms of environmental activism that people did not report that they were significantly more likely to do after the protests occurred: volunteering for or donating to an environmental cause, and participating in civil disobedience. The fact that people said that they were not more likely to participate in civil disobedience (even if they were more likely to participate in peaceful protests) suggests that people may not have been willing to do the sorts of things that Just Stop Oil or Extinction Rebellion were shown doing in the media. Rather, they were willing to participate in less disruptive activism, which is in line with what is termed the positive radical flank effect, where more radical actors can increase mobilisation or support for more moderate actors (see yet unpublished research that supports this in the context of animal rights and environmentalism).
The survey responses suggest that a significant portion of the British public became aware of Just Stop Oil, with only 38% of respondents saying that they had heard nothing at all about the climate protests by Just Stop Oil, although it is worth noting that only 5.3% said that they knew ‘a great deal’ about the climate protests that were taking place.
This research should not be seen as definitive or strong evidence for the claim that protests in general make people more likely to self-report that they are likely to engage in activism in the near future. Although it is true that we did find a marginally significant change in the number of people reporting that they were more willing to engage in climate activism, we do not believe that this research should be generalised to conclude that climate protests will necessarily lead to more people reporting that they are likely to engage in climate activism in the near future.
Value-Action Gap
If it is the case that respondents’ self-proclaimed likelihood of engaging in environmental activism did increase on account of these protests, it remains unclear whether their own perception of their likelihood of engaging in activism is actually accurate. There is widespread understanding within the field of environmental psychology that the actions of people can be different to their beliefs, commonly referred to as the Value-Action Gap (Flynn et al. 2007). In our case, we only measure the willingness of people to take part in climate activism, not whether they actually increase the amount of activism they take part in. Hence, it’s possible that despite people having increased willingness to engage with climate activism, very little behaviour changes (Castiglione et al. 2022). In future work, one could build upon this by using additional questions, possibly after a long follow-up period, asking about the number of times they have actually taken part in some of the actions we suggested in the past 6-12 months.
Climate-related external events
There were other climate-related events that took place during the survey period that may have influenced respondents’ views, namely the report by the Intergovernmental Panel on Climate Change (IPCC) Working Group Three on the mitigation of climate change released on the 4th of April and the British Energy Security Strategy that was released on the 6th of April, both before our second and third surveys. Hence, one could plausibly claim that these were the key reasons that willingness to engage in climate activism increased over our time-period.
If it was the case that the IPCC report and British Energy Security Strategy were the main drivers of shifting public opinion, we would have expected to see a larger increase in our measured variables between the Before (March 29th) to During (April 9th) surveys, shortly after the release of both reports, relative to changes in opinion when comparing our During (April 9th) to After (April 19th) survey.
Of our three climate activism variables that increased from our Before to After period, only the variable “willingness to engage with your local MP” experienced a statistically significant shift from the Before (March 29th) to During (April 9th) period. The other two variables, “willingness to talk to friends and family about climate change” and “willingness to attend a legal climate protest” showed only statistically significant increases from the During (April 9th) to After (April 19th) period or the overall period respectively.
From these findings, one could infer that whilst climate protests and related media coverage were the main contributing factors to increases in two of our three statistically significant variables, the impact of the IPCC report and British Energy Security Strategy is not easy to disentangle for the shift in “willingness to engage with MPs”. As there were both large climate protests by Extinction Rebellion and Just Stop Oil, with associated media coverage, in the period from March 29th to April 9th, in addition to the external events mentioned, causality is harder to attribute for this particular variable.
As no major climate-related events that we know of occurred between the 9th of April and the 19th of April, besides Just Stop Oil protests and related media coverage, we suggest that the increase in overall willingness to take part in climate activism, willingness to talk to friends and family, and willingness to attend a legal climate protest are largely attributable to these protests and related coverage.
Additional limitations can be seen in our Appendix.
The disruptive protests by Just Stop Oil, as well as those by Extinction Rebellion, appear to not have caused any significant backfire or negative effect on public opinion on climate change. This seems highly relevant, as common criticisms of disruptive protest is that they turn people away from supporting an issue, whereas our evidence doesn’t support this hypothesis.
Additionally, the protests may have resulted in a small increase in the number of people willing to engage in environmental activism. That being said, this should only be taken as prima facie evidence that the protests by Just Stop Oil had a weak effect on the number of people saying that they were likely to take part in environmental activism, as it is difficult to ensure that we are seeing direct causal effects when we are looking at a before-and-after study in the real world. Similarly, we only found marginally statistically significant results after adjusting for the fact that we tested multiple hypotheses, so it is possible (although we believe it to be unlikely) that the change here is merely noise. It is worth mentioning that the effect size was small (Cohen's d < 0.2), which is perhaps to be expected given the extent to which coverage of the protests was overshadowed by developments in Ukraine, although this was not something that we had anticipated so strongly prior to the surveys being conducted. However, this potentially still amounts to an additional 1.7 million people in the UK who were willing to take part in some form of climate activism, which is a non-negligible result in our view.
In brief, here are some of the pieces of additional work we’re planning on doing to strengthen this piece:
If you have specific questions or want to talk more about our research, feel free to contact James at [email protected]. If you’re interested in funding our research or curious to hear more about our future plans, please contact James here.
Thanks to Ben Kenward especially for constant help throughout designing this survey and to Moritz Stumpe, Andrea Polanco, En-QI Teo, William McAuliffe, Ren Springlea, Justis Mills, and Jo Anderson for feedback on our analysis and identifying avenues for further work. Any mistakes and errors are our own.
Agnone, J. (2007). Amplifying Public Opinion: The Policy Impact of the U.S. Environmental Movement, Social Forces, Volume 85, Issue 4, Pages 1593–1620, https://doi.org/10.1353/sof.2007.0059
Bugden, D. (2020). Does Climate Protest Work? Partisanship, Protest, and Sentiment Pools. Socius: Sociological Research for a Dynamic World, 6, 237802312092594. https://doi.org/10.1177/2378023120925949
Burstein, P. (2003). The Impact of Public Opinion on Public Policy: A Review and an Agenda. Political Research Quarterly, 56(1), 29–40. https://doi.org/10.1177/106591290305600103
Castiglione, A., Brick C., Holden S., Miles-Urdan E.and Aron A.R. (2022), Discovering the psychological building blocks underlying climate action—a longitudinal study of real-world activism, R. Soc. open sci. http://doi.org/10.1098/rsos.210006
Feinberg, M., Willer, R., & Kovacheff, C. (2020). The activist’s dilemma: Extreme protest actions reduce popular support for social movements. Journal of Personality and Social Psychology, 119(5), 1086–1111. https://doi.org/10.1037/pspi0000230
Flynn, R., Bellaby, P., & Ricci, M. (2009). The ‘Value-Action Gap’ in Public Attitudes towards Sustainable Energy: The Case of Hydrogen Energy. The Sociological Review, 57(2_suppl), 159–180. https://doi.org/10.1111/j.1467-954X.2010.01891.x
Kenward, B., Brick, C. (Forthcoming). Disruptive environmental activism in London increased environmental concern and support for such activism.
Polanco, A., Parry, J., Anderson, J. (2022). Planting Seeds: The Impact Of Diet & Different Animal Advocacy Tactics. Faunalytics. https://faunalytics.org/relative-effectiveness/
A small note is that although there was no crossover between the sample for the first survey and the second survey, there were approximately 130 people who participated in the third survey who also participated in either the first or the second survey. This is discussed further in the Appendix.
Although the change in this aggregated metric was significant prior to making the adjustment (p = 0.02), the adjusted p-value does not reach the threshold for significance (p = 0.09) after using the Benjamini–Hochberg procedure to adjust for the expected number of Type I Errors.
The change from before to after the protests had a Cohen’s d of 0.07, smaller than we had anticipated prior to carrying out the surveys, with the reasons why explored in the Discussion section.
Three weeks into motorway climate change protests, public opposition has only grown | YouGov. (2021). YouGov. Retrieved 7 April 2022, from https://yougov.co.uk/topics/politics/articles-reports/2021/10/08/three-weeks-motorway-climate-change-protests-publi
Thanks, interesting topic and glad you looked into this! (Just read the summary and skimmed the rest.) My spontaneous reaction to the results was that only days after the protest might be a little too soon to observe a backlash?
Thanks, good point! I agree that it's possible that a backlash could occur a while after the disruptive protests actually took place. That being said, it seems likely (at least to me) that if it were the case that these protests were going to lead to people becoming less supportive of climate policies, there would have been at least some evidence of the backlash in the survey data at the height of, or in the immediate aftermath of, the disruption.
Initially, we had planned only to do two surveys, but decided on commissioning a third when it became clear that JSO were receiving additional media coverage and we wanted to make sure that we captured any impact of the continued media coverage. The protests began on the 1st of April, and our third survey took place on the 19th of April, at which point most survey respondents had heard about the protests (or at least claimed that they had), so our hope is that if there had been a backlash we would've picked up on it. That being said, you're right that if there was a backlash or a change in peoples' views on climate policies that took place a few weeks after the protests (or at an even later point), our surveys wouldn't have picked up on it.