The online home for Greg Leedberg, since 1995.

Saturday, March 18, 2006

The Usefulness of Polls

I hate polls.

That said, I guess you know where this post is going. The usefulness of polls? Not much.

The modern media seems to love polls. Any big subject that comes up, CNN and Fox News will put up poll results saying that "60%" are for it, "20%" are against it, and "20%" are undecided. Great. Now, how has my life been enriched by this knowledge? Presidential approval polls sometimes say that America doesn't approve of the President. Does that mean he'll quit? Be impeached? Sometimes approval polls show that America does support the President. Does that mean he'll start to do no wrong? Be assured of a second term? There's really no conclusions one can draw from polls like this, other than that Fox News needs another "Fox News Alert". But why, really, do I hate polls? Lots of reasons...

One primary reason requires you to look into the head of an individual poll-taker. Imagine it, you're sitting at home, browsing the web or whatever. You get a call from a pollster, they ask you your opinion (generally, approve/disapprove) on some political issue. Just sitting there, alone in your house, you may not have a generally strong opinion in either direction -- it's a complex issue (i.e., "Presidential approval"). Whatever answer you give, it may very well be just a half-hearted answer primarily given just to give an opinion and get back to browsing the web. Now, multiply this half-heartedness by the thousands of poll takers that are surveyed. Suddenly, you have "clear cut" results -- based off of a thousand fuzzy responses. The fuzziness of each individual response compounds to build a conclusion that likely is quite meaningless itself -- but of course it is presented as though it's a new scientific finding.

Another problem with polls is that they are frequently based on highly complex issues that only people with relatively deep knowledge of the relevant domain can really answer -- but the question is posed to people who may have no idea at all about the subtleties involved. An example: Fox News did a recent poll regarding whether or not Usama Bin Laden would be caught. Luckily, the majority said that Bin Laden would be caught -- that sure makes me sleep better at night. But, what exactly did this "majority" base this on? These people are not military commanders out in the field conducting the operation. They aren't intelligence officers collecting data on Bin Laden's whereabouts. They are people who have absolutely no knowledge of the day-to-day events that go on in the search for the world's most notorious terrorist. This is nothing against the poll takers, but I really don't care what their opinion on a matter such as this is, because it is literally based on nothing.

Additionally, the polls themselves are set up to not really provide any information. The very questions posed to poll takers are designed to take a very complex issue, and boil it down to a handful of choices. A frequent poll question is, "how is the war in Iraq going?" That's quite a loaded question, but respondents frequently just pick something along the lines of "good" to "bad". It's so complex, though, that these responses produce more questions than they answer. Did the overthrow of the Iraqi government go well, or the rebuilding process? Is it that the trial against Hussein is going badly, or is it that there's too much money going into the war? Likely lots of people have mixed feelings on the different facets of the war, but they are given just a narrow set of responses to give to a very loaded question. This lack of options produces a narrow view on the subject, which prevents really interesting and lively debate to occur. And, of course, means that the polls don't tell us anything useful.

Lastly, the big problem is that there's this underlying expectation that the government should follow the polls on specific issues. The majority of the country is against abortion? Outlaw it! The country is for stem cell research? Fund it! We seem to lose sight that these polls are based on the opinions of people who don't necessarily know all of the subtle issues within a bigger issue. We elect government officials because we want them to know more about the important issues than we do ourselves, and then make informed decisions for us. If we just want them to follow polls, why have a representative government at all? You should make sure you pick a candidate that shares your values, and that you think will do a good job making decisions, and then let them make those decisions.

People (and so also polls) are fickle. Political decisions shouldn't be. The average person doesn't know a lot about each complex issue that comes along, and so polls on the average don't necessarily reflect an informed opinion. In isolation, a person is not inclined to put a huge amount of thought into a poll to begin with, so these uninformed conclusions are half-hearted, at best. The usefulness of polls? Hmm, not much.

Labels: ,