The UK public increasingly feel people need to be more sensitive in how they talk to those from different backgrounds, with a third (35%) now holding this view – up from a quarter (26%) at the end of 2020, according to a new study by the Policy Institute at King’s College London and Ipsos in the UK.

The shift in views means the country is now evenly split on whether people are too easily offended (35%) or whether they should be more sensitive (35%).

But at the same time, the public are still more likely to think freedom of expression (38%), rather than freedom from threatening or abusive opinions (14%), is most threatened in the UK today – although a third (32%) say both are equally under threat.

The divide in opinion on causing offence is reflected in views on how to respond to specific free speech issues – for example, 37% support cancelling a comedian’s TV show for using language that is offensive to people from minority groups, while another 37% are against.

However, there is a clearer consensus when it comes to other cases, such as banning football supporters who boo players taking the knee, which a plurality support (46% vs 28% oppose), or a police force firing an officer for offensive social media posts made when they were younger, which half oppose (52% vs 21% in support).

The findings are part of a series of studies updating on research carried out in 2020. The series, which is informed by a nationally representative survey of nearly 3,000 people, provides an in-depth assessment of the “culture war” debate in the UK.

Belonging to certain groups increases the likelihood of thinking people take offence too easily

Using statistical methods, it is possible to estimate how much more likely one section of the population is than another to believe that people are too easily offended, while controlling for other characteristics. This gives a clearer picture of how being a member of a particular group increases the likelihood of a certain outcome.

For example, Leave voters are 4.5 times more likely than Remain voters to believe people are too easily offended, while men are 3.3 times more likely than women to feel this way.

Race, sexuality, trans issues and gender identity are the issues people feel need to be discussed most sensitively, while seeing a difference between offensive and abusive speech

The public are most likely to say people should be careful not to offend others when discussing race (50%), transgender issues (41%), sexuality (39%), and gender identity (39%), while the public are comparatively more relaxed about other potentially controversial topics – for example, 20% think people should be careful not to offend others when discussing the British empire.

The extent to which the public feel sensitivity is needed on certain issues also depends on the type of speech being considered, with greater concern expressed over views that are threatening or abusive. For example, 39% think people should be careful not to express offensive views about sexuality, but this rises to 51% when it comes to threatening or abusive views.

Support for action affecting free speech depends on the specific context and issues at stake – from comedy and sport to academia and the world of work

Whether the public support or oppose actions that would potentially impact free speech depends on the specific issues at stake.

The public are relatively divided on the response to some scenarios:

• A third (36%) say they support a private company firing an employee for being a member of a political party which expresses offensive views, with virtually the same proportion (34%) against.
• 37% support a TV network or streaming platform taking down a comedian’s show for using language that is offensive to people from minority groups, while another 37% are against.

But in other cases, there is a clearer consensus (although still not always a majority on one side or the other):

• Around two-thirds (64%) oppose using physical violence to prevent hate speech, compared with 14% who are in favour.
• Half (52%) are against a police force firing an officer for offensive social media posts written when they were younger and before they joined the force, while a fifth (21%) support such action.
• Four in 10 (42%) support colleagues trying to get a co-worker fired for making sexist jokes about women, versus three in 10 (29%) who are opposed.
• The public are more likely to oppose (44%) than support (25%) organisers cancelling a talk from an academic over that academic’s views towards trans women.
• 46% are in favour of banning football supporters who boo players who take the knee from attending matches, compared with 28% who are against.

There are also big differences in views across groups within the population. For example, 61% of 16- to 24-year-olds support firing a co-worker for sexist jokes, compared with 35% of those aged 55 and above. And 2019 Labour voters (49%) are much more likely than Conservative voters (27%) to support cancelling a comedian’s show over offensive language.

The UK is made up of five groups with distinct attitudes to free speech

Through statistical analysis, it is possible to identify five different groups within the population, each with distinct attitudes towards freedom of expression:

“Free speech fighters” (12% of UK)

With men representing eight in 10 of this group – the highest of any – they are the most concerned about freedom of expression, with a very strong belief that people are too easily offended and very little support for actions that impinge on free speech.

“Free speech-concerned” (28%)

Mostly concerned about freedom of expression over freedom from harm, with some worry about people being too easily offended. They feel that, by and large, individuals and public figures should be able to say what they want, and they do not have much support for actions that affect free speech.

“Sensitive non-interventionists” (22%)

With two-thirds women – the joint-highest proportion among the different groups – they are concerned about both freedom of expression and freedom from harm, with a strong belief that people should be more sensitive in the way they talk. However, they express little support for actions that would impact free speech.

“Sensitive interventionists” (20%)

Share the concerns and beliefs of sensitive non-interventionists but diverge in also supporting a number of actions that impinge on free speech.

“Freedom-from-harm fighters” (18%)

Most concerned about freedom from harm and least concerned about freedom of expression, with a very strong belief that people should be more sensitive in the way they talk and that private individuals and public figures should be careful not to use threatening, abusive or offensive language. They support many actions impacting speech, and two-thirds are women.

Professor Bobby Duffy, Director of the Policy Institute at King’s College London, said:

“Freedom of speech has become a key focus in politics and the media - and our study shows why it is a vital dividing line, with very strong views on different sides of the debate. At one end of the spectrum, one in eight in the UK are ‘free speech fighters’, who have a very strong belief that people are too easily offended - while at the other end, one in five are ‘freedom-from-harm fighters’, who believe people should be more sensitive in how they talk, and support actions to reduce offense.

“Across the population as a whole, we’re also often very evenly divided on taking actions that affect free speech: for example, we’re split down the middle on whether videos of a comedian’s show should be taken down if it uses language that is offensive to minorities.

“But public opinion also reflects the complexity of the issues involved. The public distinguish between issues, and see it as much more acceptable to risk offence on subjects like Brexit than race or gender identity. They also distinguish between speech that causes ‘offence’ and ‘threatening or abusive’ speech.

“This is not a simple battle between two sides of a ‘culture war’, but a complicated issue that needs careful discussion.”

Gideon Skinner, Head of Political Research at Ipsos UK, said:

”Attitudes towards freedom of speech versus the need to be sensitive towards others remain a key touchpoint in the UK, and this latest research shows that views remain divided (although with a movement towards the belief that we need to be thoughtful of others in the way we talk). Our analysis also shows that, as with culture wars more broadly, there is a range of groups within the population, with different stances towards free speech. Whilst there are those at the extremes, we should not forget that many have more balanced views too.

“This a complex subject – people draw distinctions depending on the context, topic, and individuals involved, as well as whether the speech is offensive or abusive. There are clearly some issues that people believe need to be handled particularly sensitively, especially around race and gender and sexual identity, although less so on political topics like Brexit and climate change. But nevertheless there is little sign of support for a blanket use of actions that impinge on free speech, even though some think it can be appropriate in individual cases.”

Survey details
Ipsos interviewed online a representative UK sample of 2,834 adults aged 16+ between 26 November and 2 December 2020 and 2,931 adults aged 16+ between 13 and 19 January 2022. This data has been collected by the Ipsos UK KnowledgePanel, an online random probability panel which provides gold standard insights into the UK population, by providing bigger sample sizes via the most rigorous research methods. Data are weighted by age, gender, region, Index of Multiple Deprivation quintile, education, ethnicity and number of adults in the household in order to reflect the profile of the UK population. All polls are subject to a range of potential sources of error.

How the five groups within the population were identified
Latent Class Analysis (LCA) is a statistical technique that uses different response patterns in data to classify people into relatively different groups. Each person is assigned to a class with an attached probability of being a member of that class. The outcome is the identification of subgroups of individuals who are similar to each other and distinct from those in other classes.