The very nature of public opinion polls is that they are not strictly private. By definition, the questions will be seen by 2,000-or-so ordinary members of the public, and there is nothing to stop them telling their friends, taking screen shots, and sticking the questions up online. In the social media era, this sometimes leads to people getting upset about the questions – whether they are fair, or whether it is even legitimate to ask about a particular subject. This happened to us recently with a question we ran for an academic asking about the voting franchise, and whether prisoners or benefit claimants should be allowed to vote.
In that case, part of the criticism seem to be based on a misapprehension that the questions polling companies ask reflect the views that polling companies – or their owners – hold. While we do sometimes run some questions for publicity, the vast majority of the work we do is for paying clients; this is, after all, how polling companies make their money. So, the answer to the question "Why are YouGov asking that question?" is almost always “because a client has commissioned us to ask it.”
If you want to know why clients have commissioned it, then the best answer I can give is to explain who political pollsters' clients normally are. People often assume that we spend most of our time doing polling for newspapers and political parties. In reality, we do comparatively little for either. The bulk of our "political" work is actually for university academics, think tanks, public affairs companies, and charities. If you see a question about contentious or extreme views – or opinions you find offensive – then it's a fairly comfortable bet that it has either been conducted for an academic studying the issue or for a charity campaigning against that viewpoint.
Charities campaigning against racism or prejudice will often use opinion polling to measure how widespread those views are – both as a way of highlighting the issue and also to track their progress in changing public opinion. Equally the impact of populist parties in recent politics means that in recent years a lot of political academics have been commissioning polls that have included questions on attitudes that some might see as authoritarian or racist.
For example, in the past we've asked about transphobia and homophobia for LGBT charities, we've asked about anti-Semitism for those campaigning against anti-Semitism, we've asked about racist views for anti-racist campaigns. We've also asked about these issues for academics studying authoritarian populism.
That brings us to the fundamental question: even if it's for a client, is it okay to ask about these sorts of subjects? Well, as pollsters we have a duty to make sure that questions are asked in a fair and unbiased way, but assuming that they are asked appropriately we do not shy away from asking about difficult topics. If a charity wants to measure what proportion of people hold prejudiced views on sexuality, or if an academic wants data on whether or not racist attitudes in Britain are falling, the only way to get solid data is to commission polling asking if people hold these views.
The only way to understand public opinion – including those parts of it that many may find unpleasant – is to ask about it. If you want academics to be able to study and understand extremism and prejudice, they need to be able to collect data on it. If you want charities and campaigning groups to be able to effectively fight against views you oppose, then they need to be able to research and understand what they are fighting against.
So, next time you see a question about a view you find offensive, the odds are that the people looking to measuring that view are just as opposed to it as you are.