A day before the New Hampshire Primary, opinions polls indicated that Barack Obama would beat Hillary Clinton by a double-digit margin. We all know what actually happened. Why did opinion polls get it so wrong?

I came across this thought-provoking article on Slate that suggests that opinion-poll results may themselves affect the voters one way or the other. This is a reason­able surmise, but the question is, which way does it swing the voters, and can anything be done about it?

In a way, this is something that happens to any group of people trying to make a decision. If most of the members share a single opinion with convic­tion, the others may be forced to agree just so that they don’t end up feeling silly. Besides, no one wants to end up on the losing side; it feels much better to have the candi­date of your choice elected.

On the other hand, if it seems as if one candi­date has a clear-cut victory in sight, the voters may not be as inter­ested in doing their job. This is a risk that is higher for the predicted winners, since compla­cency is on their side.

There is a domino effect at play, and everyone knows it: the more support you get, the more likely it is that more people will support you. The more relevant question, I think, is: couldn’t the organi­za­tions conducting opinion polls fabri­cate the results in such a way as to tilt the real polls in one candidate’s favor? It doesn’t matter what people think; it only matters what they believe, and how they react to those beliefs.

This entry was posted on Thursday, January 10th, 2008 at 21:41 and is filed under Daily Rant. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.