The Professor Will See You Now: Island of strangers
4 min read
Lessons in political science. This week: island of strangers
When Bob Hawke, then prime minister of Australia, was told that he needed to be careful how he handled a racially sensitive issue, because of the prejudiced attitudes of some of his supporters, he is said to have replied: “Then tell me what I need to say to turn them around.”
I’ve always liked this story, both because it is a reminder that polling should not be seen as the endpoint of politics but rather as part of the process, and because it exemplifies one of the key findings of research on public opinion – that elites matter. Public attitudes are not set in stone; politicians have the power to change them.
Take, for example, a new German study into the effect of anti-immigration rhetoric, which finds that the messenger can be as important as the message. The researchers showed survey respondents roughly similar anti-immigration statements from members of either the radical right AfD or the centre-right CDU – and tested the effect these statements had on various measures of tolerance towards immigrants.
Public attitudes are not set in stone; politicians have the power to change them
They found that anti-immigration rhetoric had much more of an effect when it was made by politicians on the centre-right than when it came from those on the radical right. The views of the latter were largely shrugged off and had no effect on what the researchers called “tolerance norms” among the public. But statements from the former were more consequential; anti-immigration statements from politicians on the centre-right had an impact. “Our results,” the research team concluded, “highlight the pivotal role of mainstream politicians in enforcing or eroding democratic norms.”
Which brings us to the “island of strangers”. The psephological gods were smiling down on us when the Prime Minister gave that speech on 12 May, because it fell exactly mid-way through the polling for wave 30 of the British Election Study, which went into the field on 2 May and concluded on 22 May. Roughly half its respondents were polled before the speech, roughly half after.
Hey presto – a perfect natural experiment. (Or as the research paper puts it, “we are able to leverage the naturally occurring exogenous exposure to the news of Labour’s rhetorical shift”.) But it gets better: the British Election Study is a large survey, allowing plenty of analysis of sub-samples, and it’s a panel – that is, a survey that tracks people’s views over time. It’s polling perfection. If nothing else, the PM deserves a medal for services towards public opinion research.
Analysis of that data released recently showed that prime ministerial speeches can indeed shape public opinion. The speech increased the salience of the issue – more people said it was the most important issue facing the country – and made people overall see Labour as more right-leaning and more anti-immigration, although it had no effect on those already most opposed to immigration. It lowered evaluations of the Prime Minister and Labour’s electoral support, the latter by just over one percentage point. It appeared to have no impact on views of, or support for, Reform UK. It is not obvious this was the intended outcome.
Perhaps the most remarkable feature of this is that these effects occurred from one speech, albeit with a lot of accompanying media coverage. Can you imagine what a concerted campaign could achieve?
A caveat: one of my rules for this column is that I only report work that has been published – that is, through the peer review process and out the other side. The island of strangers research is currently still under review and is only available as a pre-print. But one of my other rules is that I can ignore my rules when it suits me, especially if I think some research is important, as this is.
Further reading: V Valentim et al, How Mainstream Politicians Erode Norms, British Journal of Political Science (2025); S Turnbull-Dugarte et al, Accommodating the Radical Right: The Electoral Costs for Social Democratic Parties (doi: 10.31235/osf.io/zbmp3_v1)
