Do a Third of Brits Endorse Violence Against Immigrants?

Yesterday evening, a rather shocking poll went viral on Twitter/X. It found that at least a third of Brits endorse violence towards immigrants. 39% of the respondents agreed that “when it comes to the refugee problem, violence is sometimes the only means that citizens have to get the attention of British politicians”. And 34% agreed that “attacks against refugee homes are sometimes necessary to make it clear to politicians that we have a refugee problem”.

These findings seem to suggest that, rather than being confined to a small minority of hooligans, troublemakers and career criminals, support for political violence is widespread. However, there are reasons for scepticism. In fact, the true portion of Britons who endorse political violence is likely much lower than a third.

As several pollsters have noted, the relevant questions suffer from two quite serious problems. First, they asked people whether they agree or disagree with particular statements. This may sound innocuous, but there is a well-known ‘acquiescence bias’ whereby respondents are more likely to agree than disagree with statements in polls.

For example, if you ask people whether they agree “the NHS needs reform more than it needs extra money”, they agree by 43% to 23%. But if you ask them whether they agree “the NHS needs extra money more than it needs reform”, they also agree – by 53% to 20%. The typical way to get around this is by giving respondents mutually exclusive options. You ask, “Which does the NHS need more: reform, or extra money?”

The second problem is that the questions only presented one side of the argument. The statement, “when it comes to the refugee problem, violence is sometimes the only means that citizens have to get the attention of British politicians” only considers one potential benefit of violence, namely getting politicians’ attention. A more neutral formulation would either not mention this benefit, or would mention a potential cost as well. Say: “violence is sometimes the only means that citizens have to get the attention of British politicians, even though it could result in innocent people dying”.

Incidentally, the company that carried out the poll has said that the questions were taken from an academic paper, and that they (the company) were asked to run them by a client.

We can see the findings are implausible by comparing them to findings from other recent polls that phrased questions more appropriately.

YouGov asked people whether “the unrest at protests” was justified or unjustified, and found that only 12% thought it was justified. (They asked separately about whether the protests themselves were justified.) Likewise, More in Common asked people whether “violent protests outside the accommodation refugees live in” would be justified or unjustified, and found that only 8% thought they would be justified.

In addition to issues with how the questions were phrased, there may also have been some inattentive responding – that is, respondents clicking through the survey without paying proper attention.

According to tables provided by the polling company, the questions were asked as part of an online poll that had several dozen in total (which may have led to boredom and thus inattentive responding). Indeed, they show that even a sizeable percentage of Liberal Democrat voters endorse violence against immigrants. Do 26% of Lib Dems really agree that “attacks against refugee homes are sometimes necessary”? I don’t buy it.

So what percentage of Britons do endorse violence against immigrants? The YouGov and More in Common polls suggest it’s closer to 10% – which seems altogether more plausible than a third.