Recently I was listening to Peter Hitchens, British journalist, political figure and brother of the late Christopher Hitchens, and he briefly spoke to the nature of opinion polls, the use of such polls and why they are misnomers. In sum he said, opinion polls don’t actually measure opinions. They are tools to influence them. This is why the questions are leading: they desire a particular response, in fact, these types of polls illicit certain results.
Here is an example:
If you want to run a story that says a majority of people support the legalization of marijuana, all that is needed to justify the headline, “majority supports marijuana legalization” is to commission a poll, asking a pool of usually about 1,000 people, whether they think legalizing pot can be beneficial, they then respond, presuming a majority will say yes, and you have your poll. You have just created the report out of thin air. This can be done with just about anything anyone has an opinion about. Think about issues like whether dogs should be allowed in public parks or should firetrucks be a particular color (in many cases, red). With precision questioning, you can get the answer you are looking for. What underlies theses polls is the reason for the existence of such a poll. The poll wasn’t asked to objectively gather information nor was this information charted and the people polled represent fairly diverse backgrounds, but its reason for being is to support a position in a story. You pick up the paper and see where, “Majority (say, 51% of people polled) of people support pot.” If the opinion poll hadn’t been taken that data wouldn’t exist.
The same goes for opinions on same-sex unions, the approval rates of the POTUS, etc. A small number of the population is polled and the headlines read, “Americans support….” or “Majority of the country disapproves….” I’ve never been asked to participate in such a poll. I know lots of people who have never answered such questions, and yet “we” support or do not support _________. It makes one question the legitimacy of the poll. Sure, if you ask folks in California, for example, about same-sex marriage, you will get a favorable response. That is not objective journalism, but of course, there is no neutrality in media. However, on the flip side, if you polled a conservative county in the country’s South, you will get a vastly different response and therefore could write a counter-story saying that the majority do not support same-sex marriage. Without randomizing those polled, ideally a large percentage of the population, such polls are subjective and biased in favor of the question(s) asked. Also, I almost forgot to mention, you have to address issues of interpretation of the poll data before you can even write the story.
It is those polls and facts that are detailed, asking tens of thousands of folks over a period of time that are more likely to be true, and usually those numbers are vastly different than the opinion poll of the day. Such polls are scientific. It is likely that the opinion poll in the paper is not. Just a tool for influence.