Anton Community Newspapers  •  132 East 2nd Street  •  Mineola, NY 11501  •  Phone: 516-747-8282  •  FAX: 516-742-5867

Michael Miller


By Michael Miller
Attention: open in a new window. PDFPrintE-mail

13 Days

1. The polling. The polling. Media obsession with polling. It’s already insufferable where I now sit on the time-space continuum, on October 17, twenty days from Election Day. Ghosts of the Future, reading this a week later, I feel sorry for you, because the polling frenzy is probably even worse where you are.

2. The media loves polls because reporting on them is cheap and easy. The work is all done for them. It fills space during lulls in which there’s nothing much more to say or write. Very quickly, the polling itself becomes the story, crowding out true analysis and oversight of important issues.

3. I’ve written about polling theory and methodology quite a bit over the years. There’s still some very questionable stuff being repeated over and over, especially on the statewide and regional level. Someone has to explain to me how 95 percent of any group supposedly has a meaningful opinion on a State Senate campaign or on a Governor when not nearly that many New Yorkers even know that Albany is the state’s capital city. Stuff like that. I’m sure the people interviewed really did give those answers, but why, and what does it really mean?

4. In a nutshell, here is what we all need to remember from this point on, through Election Day “exit” polls, post-election polls and the polling about the next four general elections, which will start around November 7. Darn.

5. When the talking head is blathering on about a poll result of 47 percent to 47 percent, it doesn’t mean that at all. The 47 percent actually represents a probability range based mostly on the number of people randomly interviewed, or “sample size.” This range (the “margin of error”) is typically between 2.7 and 4.9 percentage points.

6. That 47 percent is actually, let’s say, 44 percent to 50 percent, depending on the sample size. Just as importantly, almost all published political polls are designed to a “95 percent confidence level.” This means that you will get results within that range, 44 percent to 50 percent, 19 out of 20 times you conduct the same survey, all things being equal.

7. To get unbiased results, everything depends on quality interviews of truly random, representative pools of people. That’s getting harder and harder, which is why some polling organizations are doing door-to-door surveys or inventing new ways to poll online.

8. To bring it all together, if one poll says a race is 47 percent to 47 percent and another of similar design says it is 48 percent to 46 percent, they likely mean exactly the same thing. If another poll is an “outlier” (it’s results don’t fit with the results of other polls), it may not mean anything. About one time in twenty, the results will not be in the expected range. It happens.

9. And it especially becomes obvious when you have a dozen national polling organizations releasing results in any given week, and as many as two dozen more statewide or regional polls mixed in. Some of them are outliers, statistically insignificant. Dishonest campaigns and media organizations often seize on them for their own purposes. Don’t be fooled. Don’t be freaked out.

10. At this point, the persuasion part of any campaign is mostly over. By the time this is published, one in five readers will have already voted.

11. It’s late in the political season. Most people, even if they don’t want to admit it, really know if they’re going to vote or not and who they’re going to vote for. If someone really can’t make up their minds between any two candidates by now, there’s a very good chance that they’re not going to vote for that office, or at all.

12. There’s a hierarchy of persuasion. The higher the office, the less anyone can persuade someone else about anything. People who might actually appreciate knowing that it’s important to you for Water Commissioner Joe to be re-elected feel they can make their own minds about U.S. Senators and Presidents.

13. So the best thing to do for the next 13 days is to take a gigantic Chill Pill and relax. It’s all about turnout now, and voting machines, and lawsuits.

Michael Miller is a freelance writer, designer and strategic consultant who has worked in state and local government. Email: