Manipulating Opinion.

I find it odd that people put so much faith in opinion polls, especially when it’s apparent that the results so often support the interests of the people commissioning them. The reason for that is simple enough: too many of them aren’t conducted properly.

Opinion polls are useful as long as the core sample is genuinely representative of the general view, and as long as the right questions are asked. Too often they’re not; they’re carefully manipulated to give the desired result. One example:

Let’s suppose a debate is raging over whether hunting with dogs should be banned. On one side you have an organisation in favour of a ban, so they conduct a poll in large towns and cities where there’s no tradition of hunting. They ask the question ‘Do you think hunting with dogs is cruel?’ The people they talk to have no vested interest in hunting, so most of them answer the question with an obvious ‘yes.’ Another organisation, which largely exists to represent the interests of the landed gentry, is against the ban. They conduct their poll in small towns and villages in rural areas where hunting has long been a tradition, and ask the question ‘Do you think hunting should be banned?’ People don’t like losing their traditions, whether they’re justifiable or not, so the majority predictably answer ‘no.’ Voila! Both sides claim 75% support for their view and insist that democracy is on their side. It’s all nonsense, but those who want to believe it now feel vindicated. And this isn’t theoretical. It happened in Britain a few years ago.

The most meaningless polls of all, of course, are the ‘open’ type conducted by the media and on the internet. These are the type in which a question is thrown out to a general audience with an invitation to vote. The problem here is that most periodicals and websites attract a typical reader with a partisan leaning, and so the views that come in are usually only given by people with a strong opinion in a particular direction. They’re a million miles from representing the consensus, and are completely meaningless. Yet people still believe them, and no doubt plenty of non-aligned people have their opinions swayed by them.

The basic point to make here is that opinion polls should be treated with mistrust unless certain facts are known: how and where the poll was conducted, how big the core sample was, and exactly what questions were asked. All too often these facts are not disclosed, because all too often opinion polls are about manipulation, not information.