Presidential Preference Polls Are Like Reality TV—Fiction


2016-09-19 – As frightening as they are, I pay no attention to recent polls showing Donald Trump catching or passing Hillary Clinton. By the same token, I don’t pay attention when the polls show Hillary up. The media report their polls as if they somehow reveal the state of the contest.

They don’t. Here’s why.

Let’s take the numbers first. Assume that underlying the polls is an even split 45-45, with 10 percent undecided. This is probably a false assumption (given what we are about to discuss), but it will illustrate the 3 percent “margin of error.” A poll could report this as 48-42 in favor of Hillary one day and 48-42 in Trump’s favor a week later. The headline would be “Trump gains 6 points,” but, in fact, nothing has happened! It’s all within the margin of error. The nonsense gets even worse when we realize that the 10 percent undecided factor is significantly larger than the margin of error. And the 10 percent undecided could be 13 percent (or 7 percent). The margin of error applies to all the data, including the undecideds.

This means that a Hillary landslide could be within the margin of error of a poll predicting a Trump landslide. And a Trump landslide could be within the margin of error of a poll predicting a Hillary landslide. How useful is that?

Well, it’s useful to the media who are trying to stimulate interest in their coverage. If they get it right, their prestige is enhanced. If they get it wrong, no one remembers. I even remember polls from 1972 that gave hope that George McGovern could beat Richard Nixon. Nixon won in a landslide, carrying 49 states.

That’s one reason the polls are meaningless to me. Here’s another.

We don’t elect presidents based on the popular vote. We have electoral votes. So the distribution of votes is important. Sometimes we’re told this about the polls, sometimes not. When polls are broken down by states, however, the sample size per state is reduced. That means the margin of error is greater.

Here’s another.

Did it ever occur to you that people lie to pollsters? Especially when people claim to be unhappy with both candidates. If you want to cast a protest vote, but don’t really want to ruin the country, you might tell pollsters you’re going to vote for a third party—you want to make the candidates nervous and drive down their “totals” in a way that has little or no impact on the real results.

More uncertainty in the polls.

So I try not to pay attention. It’s pretty hard. The media is relentless in their reporting of these meaningless figures. And they are frightening when they go the wrong way. Especially this year. Polls and scandals are their stock in trade precisely because they are frightening.

But they bear the same connection to the truth as reality TV.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s